Final Report

Methodology to Assess Soil,Hydrologic, and Site Parameters that

Affect Wetland Restoration

Prepared By

M.J. Vepraskas, J.G. White, R.L. Huffman,G.P. Fernandez, R.W. Skaggs, B. Lees,

C.W. Zanner, J.M. Stucky, J.D.Gregory,W.S. Broome, J.M. Ewing, R.P. Szuch,S. Luginbhul, and G.S. Kreiser

North Carolina State UniversityDepartment Soil Science, Botany, andAgricultural Engineering, and Forestry

Box 7619Raleigh, NC 27695-7619

June 2005

Technical Documentation Page1. Report No.

FHWA/NC/2003-062. Government Accession No. 3. Recipient’s Catalog No.

4. Title and SubtitleMethodology to Assess Soil, Hydrologic, and Site Parameters thatAffect Wetland Restoration

5. Report DateJune 2005

6. Performing Organization Code

7. Author(s)M. J. Vepraskas et al.

8. Performing Organization Report No.

9. Performing Organization Name and AddressNorth Carolina State UniversityDep. of Soil ScienceBox 7619Raleigh, NC 27695-7619

10. Work Unit No. (TRAIS)

11. Contract or Grant No.

12. Sponsoring Agency Name and AddressUS Department of TransportationResearch and Special Programs Administration400 7 th Street, SW

13. Type of Report and Period CoveredFinal ReportJuly 1, 2000 to June 30, 2003

Washington, DC 20590-0001 Sponsoring Agency Code2001-09

13. Supplementary Notes:Supported by a grant from the US Department of Transportation and the North Carolina Department ofTransportation through the Center for Transportation and the Environment, NC State University.

16. Abstract

Juniper Bay is a 750 acre Carolina Bay that was purchased by the NC Department of Transportation forwetland restoration. This report summarizes the current condition of the site and reviews its history,geology, hydrology, soils, and also provides data on the target or reference areas that the Bay will berestored to. Juniper Bay was converted to agriculture in the 1970’s when its timber was removed and thearea was ditched, drained, and fertilized for row crop production. The fertilization increased the levels ofplant nutrients in the upper 12 in. of soil to levels approximately five times greater than those found in theoriginal soil. Despite the fertile nature of the soils, nutrient levels in drainage waters were low. A waterbudget indicated that ground water could comprise between 10 and 35% of the total water input to the Bay.A more precise estimate will require better estimates of evapotranspiration. Evaluation of restorationmethodologies indicated that in order for the site to meet the 12.5% criterion the ditches would haveto be plugged, field crowns removed, and surface roughness increased. This scenario was developedassuming that ground water inflow did not have a significant impact on water table levels. Examination ofvegetation soils in natural bays indicated that pond pine woodland is the most appropriate targetcommunity for the existing soils in Juniper Bay. Subsidence of the organic soil surface by up to 2 ft.precludes high pocosin or bay forest plant communities for Juniper Bay.

17. Key WordsNatural resources, ecosystems, environmentalquality, land use

18. Distribution StatementNo restrictions

19. Security Classif. (of this report)Unclassified

20. Security Classif. (of this page)Unclassified

21. No. of Pages312

22. Price

Form DOT F 1700.7 (8-72) Reproduction of completed page authorized

iii

DISCLAIMER

The contents of this report reflect the views of the authors, who are responsible for the facts andaccuracy of the data presented herein. This document is disseminated under the sponsorship of theU.S. Department of Transportation and North Carolina Department of Transportation in the interestof information exchange. This report don not constitute a standard, specification, or regulation. TheUS Government assumes no liability for the contents or use thereof.

ACKNOWLEGEMENTS

Support was provided by the Center for Transportation and the Environment in cooperation with theU.S. Department of Transportation and the North Carolina Department of Transportation through theInstitute for Transportation Research and Education, North Carolina State University.

We would like to thank the Technical Advisory Committee for assisting in this research. Thecommittee members include: James Hauser (chair), John Fisher, Matt Flint, Randy Griffin, ElizabethLusk, David Henderson, Lielani Paugh, Rodger Rochelle, Max Tate, Moy Biswas, Gordon Cashin,and Derry Schmidt.

From NC State University, this project was conducted by a number of graduate students andtechnicians. Alex Adams led the support effort for 3 years and without his leadership and servicethis project would not have been completed. James Cox and Bryan Roberts were also essential insupporting numerous field efforts.

The USDA also assisted the research. James Doolittle led the ground-penetrating radar activities for3 years, and was recently helped by Wesley Tuttle. Dr. Douglas Wysocki is working on thegeomorphology of Juniper Bay. Frank Rowe was invaluable in operating a drilling truck thatcollected the sediment cores.

iv

SUMMARY

Juniper Bay is a 750 acre Carolina Bay that was purchased by the NC Department of Transportationfor wetland restoration. This report summarizes the current condition of the site and reviews itshistory, geology, hydrology, soils, and also provides data on the target or reference areas that theBay will be restored to.

Juniper Bay lies on the Middle Atlantic Coastal Plain. Below its surface are alternating layers ofsands and clays to a depth of over 50 ft. A clay layer at approximately 20 ft. lies beneath the entireBay. In its natural state, the Bay probably supported plant communities that are described as pondpine woodland, bay forest, and high pocosin.

Juniper Bay was converted to agriculture in the 1970’s when its timber was removed and the areawas ditched, drained, and fertilized for row crop production. The fertilization increased the levels ofplant nutrients in the upper 12 in. of soil to levels approximately five times greater than those foundin the original soil. Despite the fertile nature of the soils, nutrient levels in drainage waters werelow. Organic soils cover approximately one-half the Bay. As a result of land clearing operationsand drainage, the surface of the organic soil has subsided approximately 24 in. over the previous 30yr.

The shallowest clay layer varied in depth from 20 in. to 10 ft. across the site and had an averagedepth of 60 in. This layer is probably not a single, continuous layer, but consists of overlapping claylayers of varying thicknesses. In the southeast corner of the Bay, no clay layer was detected in anarea 8 acres in size to a depth of 15 ft. This could be a point of leakage or groundwater inflow.

Measurements of hydraulic gradients across the Bay showed that ground water was entering the Bayto the east and west. It was leaving the Bay to the north and south. Most groundwater movementappeared to occur between a depth of 10 to 20 ft. This was generally below the depth of the lateralditches. However, it did appear some ground water was moving toward the surface of the Bay insome sections and may be entering ditches. A water budget indicated that ground water couldcomprise between 10 and 35% of the total water input to the Bay. A more precise estimate willrequire better estimates of evapotranspiration.

Long-term simulations using the hydrologic model DRAINMOD indicated that under currentconditions the six sites monitored within Juniper Bay do not satisfy the hydrologic criterion forwetland if the duration is based on 12.5% of the growing season. Evaluation of restorationmethodologies indicated that in order for the site to meet the 12.5% criterion the ditches would haveto be plugged, field crowns removed, and surface roughness increased. This scenario was developedassuming that ground water inflow did not have a significant impact on water table levels.

v

TABLE OF CONTENTS

Technical Report Documentation........................................................................................ii

Disclaimer and Acknowledgments .....................................................................................iii

Summary...............................................................................................................................iv

Chapter 1: Introduction........................................................................................................1

Chapter 2: Surficial Geology of Southern NorthCarolina................................................................................................................................13

Chapter 3: Historical Records of Agricultural Practicesthat Affected Hydrology and Soils at Juniper Bay .........................................................38

Chapter 4: Vegetation of Three Reference CarolinaBays in Relation to Soils and Hydrology...........................................................................56

Chapter 5: Chemical Soil Properties of Juniper Bay After15, 20, and 30 Years of Drainage and Agricultural Production.....................................92

Chapter 6: Physical and Morphological Characteristicsof a Carolina Bay Wetland Soils After 15, 20, and 30Years of Drainage and Agricultural Production............................................................126

Chapter 7: Estimating Primary and SecondarySubsidence in an Organic Soil 15, 20, and 30Years After Drainage ........................................................................................................163

Chapter 8: Nutrient Analysis of Drainage Watersfrom Juniper Bay ..............................................................................................................188

Chapter 9: Ground Penetrating Radar Evaluationof Juniper Bay’s Subsurface Stratigraphy......................................................................211

Chapter 10: Groundwater Hydrology of Juniper Bay, Prior to Restoration and Groundwater Hydrology of Reference Bays .................................................................................................................219

Chapter 11: Water Budget for Juniper Bay....................................................................250

Chapter 12: Hydrologic Evaluation of RestorationOptions for Juniper Bay ...................................................................................................285

Chapter 13: Principal Findings and Conclusions ............................................................304

Chapter 14: Recommendations and Implementationof Technology Transfer Plan............................................................................................307

Chapter 1

INTRODUCTION

Many wetland restoration efforts in North Carolina have failed to meet the relativelylimited restoration goals imposed by US Army Corps of Engineers (Corps) permits. Onsuch sites that have been reviewed by the principal investigators, it is obvious thatfailures result from multiple shortcomings in site assessment, identification of potentialfunctions, methodologies to restore wetland functions, and effective assessment of theprogress of functional restoration. The research proposed here is designed to addressthose shortcomings in a study of restoration success in Juniper Bay, a converted CarolinaBay depressional wetland in Robeson County, NC.

Juniper Bay was developed for agriculture several decades ago and currently has about300 ha of drained and intensively managed agricultural land that is not jurisdictionalwetland due to its status as prior converted agricultural land. The drainage system inJuniper Bay not only removes excess surface and ground water, but it directs runoff to adifferent location in the watershed than under previous natural conditions. The overallgoal of the research is to evaluate the strategy and performance of the restoration ofwetland functions in Juniper Bay and to test alternative restoration methods. Therestoration efforts will include:

• Plugging or filling the drainage ditches as necessary to restore historicalhydrologic functions and the directions and rates of surface and subsurfacerunoff

• Re-establishing the forest community in accordance with community typeslocated in the reference ecosystem

• Soil management as needed to assist in hydrologic function restoration, forestcommunity establishment, and nutrient cycling processes.

This research will evaluate whether these strategies are sufficient to restore appropriatewetland functions in Juniper Bay, and will identify other factors and methods that mustbe addressed in implementing wetland restoration in depressional wetlands that havebeen converted to agriculture.

BACKGROUND

A. Wetland Restoration Requirements – Wetland mitigation as practiced by the NorthCarolina Department of Transportation (NCDOT) is the restoration of wetlands to replacethose altered or destroyed in the course of road construction and maintenance. The typeof wetland ecosystem that must be replaced is specified by the Corps permit thatauthorized wetland alteration in a particular road project in accordance with Section 404of the Clean Water Act. To get full credit for wetland restoration efforts, Corps permitsusually specify that wetland hydrology, hydric soils, and a plant community similar to thereference ecosystem be restored. The reference ecosystem is a functioning wetlandlocated in the vicinity of the restoration site that has minimal alteration and that is judgedto represent the prior natural condition of the restoration site. However, most wetlands

2

have key hydrologic and soil characteristics, such that if the hydrology is restored tocause key soil processes to occur, then it is likely that the most important wetlandfunctions will be restored.B. Reasons Restoration Efforts can Fail – All of the principal investigators have beeninvolved with creating and restoring wetlands in both the southeastern and midwesternU.S. While restoration is simple in concept, it can be difficult to implement for a varietyof reasons. These include:

1. Variability in Soils and Sediments : In most of the natural wet flat, organic flat, anddepressional wetlands in North Carolina, slowly permeable soil or sediment layers nearthe surface that limit vertical or lateral drainage are instrumental in the maintenance ofwetland hydrology. Drainage ditches often penetrate such layers and provide subsurfaceflow connections to geologic sediments consisting of layers of sand or gravel. Suchconnections can cause water ponded in a wetland to leak out. This causes the wetland tobe drier than normal, and in extreme cases can limit establishment and growth of wetlandplants and the development of hydric soils. The locations and depths of both permeableand impermeable layers must be known prior to beginning restoration, and hydrologicrestoration methods on the site must restore the functional impact of slowly permeablelayers.

2. Regional Alteration of Hydrology: If ground water levels in areas outside therestoration site have been lowered by ditching or pumping, then these modified levelswill often affect the ground water levels within the restoration site. Regional subsurfacehydraulic gradients that are much higher than historical ones can subvert restoration ofwetland hydrology by contributing to relatively high rates of subsurface lateral flow innear-surface soil layers with relatively high hydraulic conductivity. Simply fillingditches within the site itself may not restore wetland hydrology if water is able to leak outthe wetland’s bottom or subsurface perimeter. Therefore, regional hydrology must beassessed and restoration methods must account for restoration of historical regionalsurface and subsurface hydraulic gradients.

3. Excessive Levels of Soil Nutrients : While all plants need certain nutrients to grow,the natural plant communities of Carolina Bays are adapted to soils that are acidic andcontain few nutrients. Fields used for agriculture were fertilized and limed regularly, sothe nutrient levels in the soils are high and the acidity low. When wetlands are created infields used for agriculture, the high levels of nutrients can cause undesired plants toflourish at the expense of the desired plants of the reference ecosystem. Conversely,restored areas of bottomland hardwood forest fail to grow adequately when grading andsoil alteration leave soils at the surface that contain inadequate nutrient levels.

4. Alteration of Soil Physical Properties: Wetland restoration often requires extensivegrading of the soil surface. For example, the typical restoration procedure for drainedagricultural land includes filling the ditches by pushing soil from the inter-ditch fieldsinto the ditches. That process destroys the soil profile. Top soil is pushed into theditches, subsoil remains at the surface, and the structure of the soil at the surface isseverely degraded by the soil movement and compaction that occurs. The soil remaining

3

at the surface has much lower site quality potential for plant community restoration thanthe previous agricultural field. Restoration methods must be developed (and approved bythe Corps) that can restore hydrology while minimizing such adverse impacts on the soil.

Successful restoration will require that potential limiting conditions such as thosedescribed above be identified in the planning process. Then the site will have to bemanaged during the construction phase and establishment phases to prevent limitingconditions and or apply management practices that ameliorate them.

PROBLEM NEED AND DEFINITION

Restored wetlands must perform the hydrologic, biogeochemical, and plant communityfunctions that are found in natural wetlands. Limited assessment of those functions isnormally conducted in accordance with Corps permits that specify monitoring of certainparameters of hydrology, hydric soil indicators, and plant community structure. Whilewetlands may perform many ecological functions, the likelihood of their occurrence canbe estimated by monitoring key soil and hydrologic properties. In current wetlandrestoration practice, there is often a lack of detailed pre- and post-restoration monitoringand assessment that documents the progress of the recovery of wetland functions onrestoration sites. This leads to conflicts between agencies creating wetlands and thoseregulating them, and delays issuance of permits until the regulatory agencies are certainthe restoration efforts will be successful.

RESEARCH OBJECTIVES

1. Document the variability in the properties of soils and sediments and the water tableregime across Juniper Bay and the reference bay that will affect restoration success.

2. Determine current groundwater flow paths and water table regime both inside andoutside Juniper Bay, and identify a strategy for hydrologic restoration.

3. Assess the recovery rate of key hydrologic, biogeochemical, and plant communityfunctions that are necessary for a sustainable wetland ecosystem.

4. Assess the usefulness of reference ecosystems for defining required hydrologic andsoils factors and target vegetation composition necessary for long-term restorationsuccess.

5. Identify soil chemical and physical properties and hydrologic requirements foroptimum growth of Carolina Bay vegetation.

6. Determine the effect of tree species type and diversity for achieving sustainable growthof desired vegetation and soil characteristics in the restored Carolina Bay.

7. Test different restoration methodologies.

4

LITERATURE REVIEW

Geographic Extent

Carolina Bays are oval, NW-SE oriented depressions with sand rims that are located inupland landscapes in the southeastern Coastal Plain and occasionally in the lowerPiedmont in certain areas. They range in size from a few acres to more than 7000 acres.Though most numerous in North and South Carolina, Carolina Bays have been identifiedas far south as north Georgia and as far north as Maryland and Delaware (Melton, 1938;Frey, 1950; Prouty, 1952; and Bliley and Pettry, 1979; ). Prouty (1952) estimated thetotal number of Carolina Bays at 500,000 with about 80% of that number occurring in theCarolinas. In North Carolina, bays are most numerous in the southern portion of themiddle Coastal Plain, including the counties of Bladen, Columbus, Cumberland, Hoke,Robeson, Sampson, and Scotland. However, they occur in most of the counties of thesouthern and central Coastal Plain and occasionally in the northeastern Coastal Plain.

Character of Soils

The bay floor may have organic soils or poorly drained or very poorly drained sandy toclayey mineral soils that are also found in irregularly shaped wet areas outside the bays(Table 1). The northwest ends of many bays merge imperceptibly with the surroundingupland, but the northeast, southeast, and part of the southwest rims are moderately todistinctly prominent landscape features. The prominent, nearly white, sandy rims withKureb or Wakulla soils on the southeast ends commonly rise 1-3 m above the bay bottomand usually are eolian (wind transported and deposited) sand. The rims on the northeastand southwest sides are sandy, but many of the sands are the result of normal soildevelopment on materials of the uplands as well as from eolian materials (Daniels et al.,1999).

5

Table 1. Major soils in the Carolina Bay System (Daniels et al., 1999)

Drainage ClassTexturalFamily Well Moderately

well Poorly Verypoorly Organic soils

Bay Interiors

Fine CoxvilleMcColl

Byars

Fine-loamy NorfolkNoboco

Goldsboro Rains Pantego

Coarse-loamy Woodington Torhunta

Sandy Lynn Haven MurvilleRutlege

MattamuskeetPamlico

Rims

Loamy WagramAutryville Bonneau

Sandy

WakullaLakelandCainhoyRiminiKershawKurebCentenary

ChipleyPactolus

Hydrology

The hydroperiod of Carolina Bays ranges from permanently flooded to seasonallysaturated. Due to the topographic gradient in bays, there is a soil drainage class gradientfrom excessively drained on the highest portions of the sandy rims to poorly drained orvery poorly drained in the lowest elevation portions. Most have significant areas ofjurisdictional wetlands, though some of the driest bays may have wetland in the lowerelevation portion surrounded by nonwetland area. Many Carolina Bays contain naturallakes, the largest located in Columbus and Bladen Counties, NC. Bays vary significantlyin types of connections to surface waters. Few have surface flow input, a notableexception being Lake Waccamaw in Columbus County. Some bays have surface runoffoutlets, but the majority likely do not. The types of surface outlets range from dispersedoverland flow during large rainfall events to well-developed stream channels.

Hydrologists have long theorized that the hydrology of Carolina Bays is influenced bysubsurface flow inputs and fine-textured soil or parent material layers that restrictdownward flux of stored water in the bay. Early limited studies of Carolina Bayhydroperiods showed that the hydroperiod was dominated by rainfall inputs andevaporation outputs (Sharitz and Gibbons, 1982). Only relatively recently, however,have detailed hydrology studies begun to elucidate the complex hydrology of Carolina

6

Bays and shown the complex subsurface interactions with the surrounding area (Knight etal., 1989; Newman and Schalles, 1990; Lide et al., 1995; O’ney et al., 1999). In the baysstudied by these authors, there was local depressional hydrology superimposed on theregional subsuface hydraulic gradients of the landscape in which the bay occurred. BothLide et al. (1995) and O’ney et al. (1999) found that the topography of subsurface layerswas similar to the surface topography. Sedimentary layers sloped downward from thesurrounding uplands to lows under the lower elevation portions of the bay. Hydraulicgradients into the bays resulted in subsurface flows along sandy layers overlying fine-textured layers with upward gradients into the bays during the wet season. In both bays,water accumulated in the bay during the wet season of the year and then was depletedduring the dry season. Lide et al. (1995) concluded that Thunder Bay in BarnwellCounty, SC provided significant ground water recharge during the drying period of latespring/early summer. O’ney et al. (1999) concluded that Chapel Bay in BambergCounty, SC likely provided some recharge but that drying was dominated by evaporationlosses.

Geomorphology

Stratigraphy of the lacustrine bay-fill sediment and the fossil pollen in the sedimentindicates water levels have fluctuated in the past. In most bays a series of alternatingorganic and inorganic zones can be identified, and most bays were more lake-like at onetime (Whitehead, 1965). Inorganic sediments consist of sandy or clayey loam, lenses ofgravel, sand, or iron-cemented sand, and marl and clay. The clay and silt zones representlacustrine depositional periods after bay formation, and as such did not influence waterlevels early in bay history. Original bay water levels probably reflected regionalhydrology. Surface water levels have decreased in time because bays are infilling withsediment and peat and surrounding groundwater levels are decreasing because of localstream excision (Schalles et al., 1989). Ditching and channelization, for primarilyagricultural drainage, have also lowered groundwater levels in their vicinity.

Some bays contain extensive organic deposits, while others are clay based. Significantpeat reflects a more stable hydrology with almost continuous groundwater recharge.Chemistry of water and soils in clay-based Carolina bays indicates a rainwater-dominatedsystem characteristic of perched-water settings (Schalles et al., 1989). Water levels arerelated to precipitation, but variable responses are common.

Sediments on the coastal plain were deposited by both coastal and fluvial processes (andsurfaces have been reworked by wind). Consequently, coastal plain stratigraphic unitscan consist of sands, silts, or clays, reflecting the energy present in the environment at thetime of their deposition. Surficial sands often overlie clay layers; these finer-texturedsediments perch water and may be important to maintaining the bays.

The original surface of these coastal plain sediments was most likely undulating. Watertables on broad interfluves would have been high; as stream incision progressed, watertables would become lower near interfluve edges. Undulations in the region where water

7

could pond because of poor surface or subsurface drainage resulted in bays(Kaczorowski, 1977).

Bays have been reported to form on saprolite and clayey Coastal Plain sediments (Blileyand Burney, 1988), over impervious humate (Thom, 1970; Kaczorowski, 1977), clay(Gamble et al., 1977; Schalles et al., 1989), and in poorly drained depressions in sandysurficial sediments (Bliley and Pettry, 1979). Landscape position, water tablefluctuations, and impervious layers interact to produce differences in individual bayhydrology and response to rainwater inputs. Bays are likely both recharge and dischargefeatures depending on bay water levels in relation to the regional water table (Schalles,1979). Understanding the hydrology of a particular bay requires an understanding of thenature, continuity, and depth of underlying sediments and their interactions with localhydrology.

Vegetation

The vegetative communities of Carolina Bays in the Carolinas are diverse among baysand usually complex within bays (Sharitz and Gibbons, 1982; Schafale and Weakley,1990). That diversity and complexity is related to topography, soils, hydrology, anddisturbance history. Schafale and Weakley recognize nine natural community types thatcommonly occur in Carolina Bays in North Carolina: low pocosin, high pocosin, smalldepression pocosin, pond pine woodland, peatland Atlantic white cedar forest, bay forest,cypress savanna, small depression pond, and natural lake shoreline. However, examplesof most common forest or emergent wetland vegetation types that occur in the CoastalPlain may be found in the complex vegetation mixes of Carolina Bays. All bays have avegetation gradient from the xeric communities of the sandy rims to the wetland oraquatic communities of the lowest elevation area of the bay, usually in the southeasternquadrant (Sharitz and Gibbons, 1982).

The long-term success of the Juniper Bay restoration project will be determinedpartly by the similarity of the species composition and structure of its future stable plantcommunity to that at reference Carolina bay sites. Certain key wetland functions atJuniper Bay should also be compared with functions at reference sites. In the early yearsfollowing restoration at Juniper Bay, before a stable community is attained, successionalchanges in the community structure and wetland functions should be monitored to try todetermine if they are progressing toward their desired future states. Clearly, determiningthe success of the Juniper Bay restoration requires characterization of reference sites.

Carolina bay lengths range from 50 m to 8 km (Frey, 1949). There appears to beno consistent relationship with particular geological formations or topography (Prouty,1952). Some bays remain dry nearly all of the time; some contain permanent water;others are seasonally inundated (Sharitz and Gibbons, 1982). Many bays have beendisturbed by forestry or agricultural practices while some remain in relatively pristinecondition. Fire is a natural disturbance that has probably affected all bays, but to varyingextents.

8

Against the background of extremely heterogeneous environmental influences, itis not surprising that plant community types vary widely among Carolina bays. Majorcommunity types previously reported at bays include pine forests, herbaceous marshes,shrub bogs, deciduous forests, evergreen bay forests, pond cypress swamps, prairies, andsubmerged aquaticbeds (Buell, 1946; Penfound, 1952; Whitehead and Tan, 1969; Porcher, 1966; Wharton,1978; Schalles and Shure, 1989). This vegetation variation suggests that the stablecommunity type that will eventually develop at Juniper Bay is difficult to predict, andthat several reference sites should be selected. These sites should be located close toJuniper Bay and have soils, hydrology, and disturbance histories similar to those atJuniper Bay prior to its conversion to agriculture. It is expected that the plant communitytypes and wetland functions will vary among these referencesites. The range of structural and functional variability in the plant communities of thesereference sites is the target for Juniper Bay.

Wetland Soils or Hydric Soils

Hydric soils are defined as those that are saturated, flooded, or ponded long enoughduring the growing season to develop anaerobic conditions in the upper part. Hydricsoils must be anaerobic. Many water quality functions that wetlands perform occurbecause the soils have become anaerobic and chemically reduced.

Wetland restoration efforts are successful when the soils in the restored wetlands becomeanaerobic for extended periods of time. Development of anaerobic conditions can bemonitored by analyzing water chemistry or by measuring the oxidation-reductionpotential, which is also called the redox potential (Ponnamperuma, 1972). Waterchemistry measurements are used to determine concentrations of reduced chemicalspecies such as NH4

+, and Fe(II). The presence of Fe(II) is considered proof that the soilsare anaerobic.

Redox potential measurements are electrical measurements that determine the voltagedeveloped between a Pt wire and a reference electrode buried in the soil. (Patrick et al.,1996). They are used to determine whether soils are developing the anaerobic conditionsnecessary for them to be considered as hydric soils. Redox potential measurements areprobably the single most important measurements to be made on soils to confirm that awetland restoration has been successful.

The time required to restore wetland functions is site specific. However, it has beenfound that some soils regain their hydric soil processes within years of construction(Vepraskas et al., 1999). In a study of created wetlands in the midwestern U.S.,Vepraskas et al. found that hydric soil field indicators developed within 3 yrs of wetlandcreation, and appeared to reach full development within 5 yrs. To develop the necessarylow redox potential required for anaerobic conditions, soil organic matter levelsapparently had to exceed 4% in the midwestern U.S. The hydric soil field indicatorsadopted for use throughout the U.S. have been presented in USDA-NRCS (1998).

9

ORGANIZATION OF THE REPORT

The following chapters summarize the principal findings of the research done to date.Each chapter covers a separate topic and was written by the researchers involved. Thechapters were written to be “self-contained” and can be read without referring to otherchapters. This format was selected to make reading of the report easier.

10

REFERENCES

Bliley, D. J. and D. A. Burney (1988). “Late Pleistocene climatic factors in the genesis ofa Carolina bay.” Southeastern Geology 29(2): 83-101.

Bliley, D. J. and D. E. Pettry. 1979. Carolina bays on the eastern shore of Virginia. SoilSci. Soc. Am. Jour. 43:558-564.

Buell, M.F. 1946. Jerome Bog, a peat-filled "Carolina bay." Bull. Torrey Bot. Club 73:24-33.

Daniels, R. B., S. W. Buol, H. J. Kleiss, and C. A. Ditzler. 1999. Soil systems in NorthCarolina. Tech. Bull. 314, Soil Science Dept., NC State Univ., Raleigh, NC.

Estes, J.E., M.R. Mel, and J.D. Hooper. 1977. Measuring soil moisture with an airborneimaging passive microwave radiometer. Photogrammetric Engineering andRemote Sensing 43:1273-1281.

Frey, D.G. 1949. Morphometry and hydrography of some natural lakes of the NorthCarolina coastal plain: the bay lake as a morphometric type. J. Elisha MitchellSci. Soc. 65: 1-37.

Frey, D. G. 1950. Carolina bays in relation to the North Carolina Coastal Plain. J.Elisha Mitchell Sci. Soc. 66:44-52.

Gamble, E. E., R. B. Daniels, et al. (1977). “Primary and secondary rims of CarolinaBays.” Southeastern Geology 18: 199-212.

Houser, P.R., W.J. Shuttleworth, J.S. Famiglietti, H.V. Gupta, K.H. Syed, and D.C.Goodrich. 1998. Integration of soil moisture remote sensing and hydrologicmodeling using data assimilation. Water Resour. 34 (12):3405-3420.

Kaczorowski, R. T. (1977). The Carolina Bays: a comparison with modern orientedlakes. Columbia, South Carolina, Coastal Research Division, Department ofGeology, University of South Carolina.

Knight, R. L., J. S. Bays, and F. R. Richardson. 1989. Floral composition, soil relations,and hydrology of a Carolina bay in South Carolina. p. 219-234, In R. R. Sharitzand J. W. Gibbons (eds.). Freshwater wetlands and wildlife symposium:perspectives on natural, managed, and degraded ecosystems. US Dept. of Energy,Off. of Sci. and Tech. Info., Oak Ridge, TN. Conf. 8603101, DOE SymposiumSeries No. 61.

Lide, R. F., V. G. Meentemeyer, J. E. Pinder, III, and L. M. Beatty. 1995. Hydrology ofa Carolina bay located on the upper Coastal Plain of western South Carolina.Wetlands 15(1):47-57.

11

Melton, F. A. 1938. Possible late Cretaceous origin of the Carolina “bays.” Geol. Soc.Am. Bull. 49:1954.

Newman, M. C. and J. F. Schalles. 1990. The water chemistry of Carolina bays. Archivfür Hydrobiologie 118:147-168.

O’ney, S. E., M. H. Eisenbies, and M. Miwa. 1999. Hydrologic processes in the vicinityof a Carolina bay affecting water quality: an assessment in association with ahardwood fiber farm. Unpublished progress report. USDA Forest Service,Center for Forested Wetlands Research, Charleston, SC.

Penfound, W.T. 1952. Southern swamps and marshes. Bot. Rev. 18: 413-446.

Poe, G.A and A.T. Edgerton. 1971. Determination of soil moisture content with airbornemicrowave radiometry. Summary report. U.S. National Oceanic and AtmosphericAdministration & Aerojet-General Corporation Microwave Division, El Monte,CA.

Porcher, R.D., Jr. 1966. A floristic study of the vascular plants in nine selected CarolinaBays in Berkeley County, S.C. MS thesis. Univ. SC, Columbia.

Prouty, W. F. 1952. Carolina bays and their origin. Bull. Geol. Soc. Am. 63:167-224.

Richter, J.C. 1981. Ground registration of data from an airborne multifrequencymicrowave radiometer (MFMR).USDA. Lyndon B. Johnson Space Center (JSC)# 17152, NASA/JSC. AgRISTARS . Lockheed Engineering and ManagementServices Company. National Technical Information Service, 1981. 29 p.

Schafale, M. P. and A. S. Weakley. 1990. Classification of the natural communities ofNorth Carolina: Third approximation. North Carolina Natural Heritage Program,NC Dept. of Environment, Health, and Natural Resources, Raleigh, NC.

Schalles, J. R. (1979). Comparative limnology and ecosystem analysis of Carolina bayponds on the upper coastal plain of South Carolina. Atlanta, GA, EmeryUniversity.

Schalles, J. F., R. R. Sharitz, et al. (1989). Carolina bays of the Savannah River Plant.Aiken, S. C., Savannah River Plant National Environmental Research Park.

Schalles, J.F. and D.J. Shure. 1989. Hydrology, community structure, and productivitypatterns of a dystrophic Carolina bay wetland. Ecol. Monog. 59(4): 365-385.

Schmugge,T.. 1976. Microwave radiometry for soil moisture sensing. p. 184-205 inRemote Sensing of Soil Moisture and Groundwater, Workshop Proceedings,1976.

12

Sharitz, R. R. and J. W. Gibbons. 1982. The ecology of southeastern shrub bogs(pocosins) and Carolina bays: a community profile. FWS/OBS-82/04. Divisionof Biological Services, US Fish and Wildlife Service, Washington, DC.

Thom, B. G. (1970). “Carolina bays in Horry and Marion Counties, South Carolina.”Geological Society of America Bulletin 81: 783-814.

Wharton, C.H. 1978. The natural environments of Georgia. Geol. And Water Res. Div.Res. Planning Sect., Office of Planning and Research, Ga. Dept. of Nat. Res.,Atlanta, Ga. 227 pp.

Whitehead, D. R. (1965). Palynology and Pleistocene phytogeography of unglaciatedeastern North America. In The Quaternary of the United States. H. E. Wright, Jr.and D. G. Frey. Princeton, N. J., Princeton University Press: 417-432.

Whitehead, D.R. and K.W. Tan. 1969. Modern vegetation and pollen rain in BladenCounty, North Carolina. Ecology 50: 235-248.

13

Chapter 2

SURFICIAL GEOLOGY OF SOUTHERN NORTH CAROLINA AND JUNIPERBAY

C.W. Zanner

INTRODUCTION

This chapter reviews the geologic history of the Juniper Bay and Bladen Lakes region,and then fits the Juniper Bay core stratigraphy into the stratigraphy of that region. Thesubaerial (emerged) portion of the Atlantic Coastal Plain extends from Long Island, NYto Florida. About 45% of the state of North Carolina is on the Coastal Plain (Daniels etal., 1999). Juniper Bay is in Robeson County NC, in an upland area of the Middle CoastalPlain. Robeson County is bordered by South Carolina to the SW and Bladen County NCto the NE. The Bladen Lakes reference bays are in the Cape Fear River Valley in BladenCounty.

The basement rock below the coastal plain sediments consists of a series of archesand embayments that dip to the east. The Coastal Plain is thus a wedge of marine andnear shore deposits that thicken to the east. Coastal Plain sediments are thicker in theembayments, but thinner and less complete over the arches, with major unconformities.The area of the bays has been under water many times during and since the Cretaceous(Table 1 shows the portion of the geologic time column that is relevant to thisdiscussion). Juniper Bay is located on the Cape Fear Arch, just south of the hinge. Thereference bays are on the arch north of the hinge.

Sediments on the coastal plain were deposited by both coastal and fluvialprocesses. Figure 1 is a generalized model of the Cretaceous delta to shelf lithofaciesdeveloped by Sohl and Owens (1991) for the Carolinas. Although this model wasdeveloped to explain the Cretaceous stratigraphy, transgressions and regressions after theCretaceous reproduced Sohl and Owens’ relationships along this part of the AtlanticCoast. Consequently, there is considerable lateral and vertical variation. Laterally, coastalplain stratigraphic units consist of sands, silts, or clays, reflecting the energy present inthe environment at the time of their deposition. Vertically, surface sands often overlieclay layers. The finer-textured sediments perch water and are likely an important factor inbay formation. The most recent surface sands have been reworked by wind. At any oneplace on the Coastal Plain, the stratigraphy of the below ground sediments reflects itsformer location from under deep water when the shore was to the west to subaerial whenthe shore was to the east. See Gohn (1988) and Sohl and Owens (1991) for location of theLate Mesozoic and early Cenozoic coasts. See Ward et al. (1991) for the Pliocene andEarly Pleistocene coasts.

Sediments covering the Cape Fear Arch show drainage patterns and deeplyincised river courses that were influenced by the lowering of sea level associated withglaciations. Sea level decreased 120-135 m (394-443 ft) during the most recent

14

Pleistocene glaciation which ended about 20,000 years ago (Mitrovica, 2003; Peltier,2002; Yokoyama et al., 2000). At 18,000 years ago, the great volumes of water tied up inglacial ice caused the coast to move 100-150 km (62-93 mi) east of where it is now(Riggs and Belknap, 1988). Older glacial advances that reached Kansas and Nebraska inNorth America (Andersen and Borns, 1997) suggest that in these older periods sea levelswere also lower than at present. Elevation of sea level was also controlled by geologicevents such as continued uplift of the Cape Fear Arch through the Quaternary (Owens,1988), and changes in ocean basin volume and general upwarping of the eastern NorthAmerican continent (Colquhoun et al., 1991). Changes in sea level, plus possibleincreases in effective moisture during glacial periods, marked periods of major incision.Major fluctuations in sea level inundated and reexposed the area, with inundations filingin the previously incised landscape. Sea level lowering was associated with periods ofweathering, erosion, and incision. Even those transgressive/regressive sequences that didnot reach the elevation of the bays changed base level of the local rivers and streams,which then incised or backfilled depending on proximity to the ocean.

GEOLOGY IN THE AREA OF JUNIPER BAY AND THE REFERENCE BAYSys

Walker and Coleman (1987) estimated Coastal Plain sediments to be 17%Cretaceous, 27% lower Tertiary, 26% upper Tertiary, and 30% Quaternary. In southernNorth Carolina, Cretaceous deposits are more predominant, approaching to within 100 k(62 miles) of the ocean. Figure 2A shows geology, primarily of the Cretaceous units, ofthe Juniper Bay and reference bays areas (North Carolina Geologic Survey, 1991). Themore detailed map (Figure 2B from Owens,1989) shows that most of the Cretaceous unitsare covered with more recent Pliocene sediments. Tables 2A and B provide a key to theunits shown on each map. Most surfaces have been extensively reworked by wind and inplaces buried by fluvial deposits, especially in the Cape Fear Valley.

Formations discussed here are summarized in Table 3 which shows the formationname, age, description, and the environment of deposition. Table 3 and the followingdiscussion are based on material presented in Gohn (1988), Sohl and Owens (1991),Soller (1988), Soller and Mills (1991), Ward et al., (1991).

Cretaceous units of North Carolina formed in deltaic and shelf environments, withrapid facies changes (Owens and Sohl, 1989). Figure 1 is a generalized model for thedepositional environments of the Cretaceous units in the Carolinas (Sohl and Owens,1991). Within the Cretaceous section, there are a number of cycles of sedimentationseparated by unconformities. These sediments are Upper Cretaceous in southern NorthCarolina (see Table 1). The oldest part of the section, the Cape Fear Formation (Table 3),is mainly visible in river systems. The Middendorf Formation covers the Cape Fear Arch.Both of these units occur northwest of Juniper Bay at higher elevations. Middendorf is ator close to the surface ~75 km (47 mi)northwest of Juniper Bay (Owens, 1989). Theseunits are likely represented in the subsurface of the Juniper Bay Project research area, butdeeper than what we cored. The Middendorf consists of light-colored iron-stained sandsand variably colored clays. Coarse-grained sands and pebbles are common.

15

The Black Creek Group is above the Middendorf. The Black Creek is a groupdivisible from younger to older into the Donoho Creek Formation, the Bladen Formation,and the Tar Heel Formation (Sohl and Owens, 1991). Each younger formation of theBlack Creek is found closer to the present ocean. In the region of Juniper Bay, Figure 3(from Sohl and Owens, 1991, page 201) suggests Juniper Bay and the Reference Baysoccur in the area mapped as Bladen Formation, near the boundary with the Tar HeelFormation, which extends farther inland. The Black Creek Formation consists of marinesediments of onshore, nearshore, and offshore origin with some deltaic deposits exposedon the Cape Fear Arch. Deltaic deposits are not surprisingly more common near riversystems. Uplift of the Cape Fear arch may have contributed to erosion and dissection ofthe Black Creek and other Pre-Pliocene surfaces. Figure 3 also shows the depositionalenvironments associated with the Cape Fear. Note that expression of the delta of the CapeFear has persisted since Cretaceous time. Figure 1 shows a more detailed model of deltato shelf facies that were deposited at any time when the river mouth was stable for anextended period.

The Black Creek generally consists of bedded black clay and micaceous sands.Donoho Creek is more massively bedded with abraded shells. Bladen Formation beds aremore thinly bedded without fossils. Tar Heel is thinly bedded, carbonaceous, andmicaceous, without fossils. Owens and Sohl (1989) described the Tar Heel Formation ashorizontal beds of thin black clay and light-colored micaceous sand.

The PeeDee Formation overlies the Donoho Creek Formation, but it does notextend inland to the Bay area. No Eocene, Oligocene, or Miocene sediments are found inthe area of the bays (Harris and Zullo, 1991; Snyder et al., 1991).

Dissected Cretaceous sediments were filled in by later transgressions. The LatePliocene Duplin Formation was deposited over the Upper Cretaceous strata over much ofthe Cape Fear Arch (Ward et al., 1991). This formation was deposited 3.0-3.5 millionyears ago (Cronin, 1991). Subsequent transgressions inundated smaller areas (see Wardet al., 1991). Current practice is to call this stratigraphic unit the Duplin-YorktownFormation in North Carolina (Kathleen Farrell, North Carolina Geological Survey,personal communication, 2003). The Yorktown Formation of Virginia and the DuplinFormations are stratigraphically equivalent but have different fossil assemblages. Threemembers of the Yorktown that were recognized in Virginia can be traced into NorthCarolina (Ward and Blackwelder, 1980). The Sunken Meadow member is not traceablesouth of the Neuse (Ward et al., 1991). The Rushmere and Morgarts Beach members arenot distinct south of the Neuse (Ward et al., 1991). At the Tarheel locality (BladenCounty NC) described by Ward et al., (1991) the sediment underlying the Duplin-Yorktown Formation is the deeply burrowed eroded surface of the Black Creek Group.Common mollusks in the Rushmere and Morgarts Beach members are Mulinia sp. andChesapecten sp. Although Chesapecten sp. occurs in other Members and Formations(Raysor, Chowan, Bear Bluff) discussed by Ward et al. (1991), Mulinia does not (seetables on pages 285-288 in this reference). Duplin beds are composed of medium tocoarse sand, sandy and silty clays, and shelly fossiliferous sands. Grayish-blue sandy clay

16

with a low diversity fossil assemblage buries the oldest Duplin deposits, and, in turn, isburied by 3-m (16 ft.) thick quartz sand containing an open marine molluscanassemblage.

The reference bays occur in an area between the Black River and the Cape FearRivers. The area has been extensively reworked, as the Cape Fear, which once flowed inthe area where the Black now flows, deflected to the southwest. This deflection persistedthrough the Pleistocene, preserving increasingly older terraces to the northeast. Distinctterrace breaks suggest changes in base level and climate. The effects of this ongoingsoutherly migration can be seen in the steep bluffs on the south side of the river whencompared to the north side. The more detailed maps of Owens (1988; 1989) and thepublication of Soller (1988) shows that the broad area between the present location of theCape Fear River and the Black River is a series of five terraces that are 2.75 million to10,000 years old. The reference bays are on the Penholoway Formation Terrace that is750,000 years old. This reworked surface was covered by a mixture of flood plaindeposits from sands to clays, reflecting energy of the system at the time of deposition.Wind has since extensively reworked and spread out the sandier deposits over most ofthese terrace surfaces. The Black Creek and PeeDee Formations are exposed in thevalley. A few areas of Duplin and Waccamaw Formations covering the Black Creek arealso exposed.

SURFICIAL GEOLOGY

The stepped character of the Coastal Plain (a series of terraces and scarpsdecreasing in elevation towards the Atlantic) is attributed to a downward trend in sealevel through Quaternary time (Colquhoun et al., 1991).

Elevation in and outside Juniper Bay ranges from 35-41 m (115-134 ft.). Thisplaces Juniper Bay on the Sunderland Terrace, which has an elevation of 30-47 m (100-155 ft), and is bounded by the Surrey Scarp to the southeast (Fig. 2B). The OrangeburgScarp to the northwest (Fig. 2B) is equivalent to the Coats Scarp of Daniels at al. (1999)and was created by the transgression that left the Duplin Formation sediments (Dowsettand Cronin, 1990). According to Soller and Mills (1991), the Duplin surface lies at lessthan 40 m (131 ft) in southern North Carolina. The Duplin surface should thus be at ornear the surface in the area of Juniper Bay.

The closest cross-section of Owens (1989) in the vicinity of Juniper Bay (~24 km(14 mi) distant) shows 15-20 m (50-66 ft) (of Duplin Formation below the surface. TarHeel Formation is below the Duplin south of the Bay, Bladen Formation north of theBay. The Bear Bluff Formation mapped by Owens (1989) shows Juniper Bay situatedless than 3 km (2 mi) from the high stand of this transgression. Retreat after high standsleft surfaces covered with sandy beach and barrier deposits that then could be reworkedby wind. Juniper Bay occurs between parallel NE-SW ridges that are obvious in theRobeson County Soil Survey (McCachren, 1978), in U.S.G.S. digital orthophotography(Figure 4), and in Digital Elevation Models (Figure 5). These ridges are remnants of thecoastal system that was once in this area.

17

JUNIPER BAY CORES AND THE LOCAL GEOLOGY

Table 4 gives a brief summary of 12 cores collected inside the bay. Core locations areshown in Figure 6. Stratigraphy of Cores 1, 2, 3, 6, 8, and 9 is shown in Figure 7. Theuppermost sediments were subaerial during at least four geologic periods. Woodfragments recovered from 1.8-2.4 m (6-8 ft.) depths have Holocene ages: Core 4--3720+/- 90 years before present, Core 10--8320 +/- 100 YBP, Core 6--8460 +/- 100 YBP.Other wood fragments from 2.4-3 m (8-10 ft) deep have Late Pleistocene ages: Core 12:35200 +/- 590 YBP, Core 5: 39600 +/- 1760 YBP and Core 16: 43880 +/- 1660 YBP.Each of these ages reflects a time when Juniper Bay was dry enough for trees to havebecome established. The presence of former surfaces, recognizable because of color andstructure, also indicate dry periods when soils could form. The trees that left this olderwood behind were growing in the Duplin-Yorktown Pliocene age sediments that are 3.0-3.4 million years old.

Cores that were collected to ~6 m (20 ft.) are grounded in this Duplin-YorktownFormation. This determination is based on the description of Duplin-YorktownFormation beds as being medium to coarse sands interbedded with thin gray clays andoften fossiliferous. Core 9 was noticeable because of the large number of mollusk shellsfound between 3.8 m (14 ft) and 5.3 m (20 ft). These shells have been tentativelyidentified as Mulinia and Chesapecten. Ward et al. (1991) describe these as two of thecommon genera in the Rushmere and Morgarts Beach Members of the Duplin-YorktownFormation. The rest of the Duplin fauna reflect a high energy, that is, not a deepwaterenvironment, associated with shallower water over the Cape Fear Arch. Core 9 has wellpreserved although somewhat broken shells. Core 20 has shell ghosts at 2.4 m (8 ft), Core22 at 5.5 m (18 ft), and Core 27 at 1.5 m (6 ft). The absence of these fossils and/or ghostsin the other cores is further indication of differential erosion and or deposition. Figure 7and Table 4 shows the upper sediments are a mix of sands and clays. Although thesurface elevations vary by only ~0.5 m (Figure 8A), the depth to the uppermost claytextures (Figure 8B) shows more variation. Cores 3 and 8 also indicate that over thedistance of the bay there was considerable variation in depositional environments and/orvariations in erosion patterns across the area of the Bay.

Below ~6 m (20 ft), rhythmites become common, finer textures and black colorsbecome more dominant. Finer textured rhythmites are separated by thin sand laminations.Sands are micaceous. Shells found lower in the section are extremely abraded. Thisdescription is consistent with these sediments being part of the Black Creek Group. Core19 outside the Bay has a shell hash at 14.5 m (48 ft) suggesting that this represents thelower part of the Donoho Creek (Table 3). Sohl and Owens’ map (1991) shows theBladen Formation of the Black Creek Group in the Juniper Bay Area. If thin beds and nofossils are enough to distinguish between group members, sediments under the Bay itselfare more likely those of the Bladen Formation. This description is consistent with whatwas observed: 1-2 cm (0.5 in) thick rhythmites separated by sands, very dark, with a highorganic content, suggesting deposition in a back bay setting with some tidal influence.However, Core 8 is clay almost to the surface and is massive, matching more closely theDonoho Creek description.

18

Figure 9 shows coarse sand percentages for selected cores inside and outside thebay. The top of Core 3 may be affected by spoil created when the ditches were installed.However, a slight coarsening is seen near the surface of all the cores, suggesting theincrease in coarse sand may be relative, and reflect wind or water winnowing. The coresshow an unconformity, most likely an erosional lag, 2-4 m (7-13 ft) below the surface. Insome cases, this lag corresponds to a buried soil. This is consistent with the suggestionmade above that the Bay experienced dry periods and suggests the sediment above thispoint represents bay fill. Core 28 is from the dune form to the southeast of the Bay thatcan be seen in the DEM and aerial photos. The 10 m (33 ft) of sand above the contact at28 m (32 ft) consistent with the interpretation that this is a dune form. Cores 8 and 28show increases and decreases in coarse sand that one would expect in an environmentwhere there were shifts in energy of deposition.

Figures 10A and 10B show clay percentages for cores inside and outside the bay.As is the case with the presence/absence of mollusk shells, the presence/absence ofburied soils, the depth of erosional lags, and the depth to the uppermost clayey sediments,these figures point to a variety of depositional and erosional environments in and aroundJuniper Bay. These figures have been projected to show their respective elevations, withthe hope that when corrected for elevation one could find it easier to project a surfaceacross the Bay that can be proposed as a surface of the same age. As stated in theintroduction to this piece, there is considerable lateral and vertical variation, even withinan area the size of Juniper Bay. Core 17 has very low clay but Cores 6 and 10 are up to60% clay. The presence of sand over an area with clayey sediment that can perch waterclose to the surface would be a logical requirement for the formation of a bay (and Core22 taken in the unnamed bay to the north also meets that). Core 17 seems to violate thishypothesis, but note that there was particularly poor recovery in this core. Cores 11 and14, which indicate some but a minor clay increase also experienced poor recovery. Themost likely explanation for poor recovery is that a thin but important clay layer wasunderlain by saturated sand, and the core tip once filled with >20% clay just pushed thedenser clay out of the way into the saturated sand. With better recovery, we might haveseen a more continuous clay layer across the bay.

CONCLUSIONS

Juniper Bay formed in 5-8 m (16-26 ft) of Pliocene aged Duplin-YorktownFormation sediments that are underlain by Cretaceous aged Donoho Creek and BladenFormations of the Black Creek Group. Most of the underlying sediment seems to be partof the Bladen Creek Formation, with Donoho Creek possibly remaining at Core 8. Thetopography at the top of each of the subsurface sediments is irregular, with the newersediment filling in erosional channels as it was deposited and then in turn was itselferoded after deposition. These irregular surfaces, along with coarse sand lags, buriedsoils, and radiocarbon dated wood, all indicate that cycles of exposure and burial havebeen occurring in this landscape since the Cretaceous. The sands of the last transgressionfilled in incised eroded surfaces as beaches moved to the east until they reached theirpresent location. Wind scour of these sand filled low areas created a series of “bays” of

19

varying sizes. Depth and thus size of any one bay was limited by depth to fine texturedsediments. Climate driven increases and decreases in water table levels resulted inperiods with wet bays (rise in water tables) or dry basins (lowering of water tables).Larger deeper bays produced waves with higher energy. During wet periods, wave actionhelped bays to grow and capture adjacent basins, analogous to stream capture. Thehorizontal distance between less erodible finer textured sediments controlled the ultimatesize of any bay.

20

REFERENCES

Andersen, B.G. and Borns, H.W., Jr., 1997. The Ice Age world; an introduction toquaternary history and research with emphasis on North America and NorthernEurope during the last 2.5 million years. Scandinavian University Press, Oslo.

Colquhoun, D.J., Johnson, G.H., Peebles, P.C., Huddlestun, P.F. and Scott, T., 1991.Quaternary geology of the Atlantic Coastal Plain. In: R.B. Morrison (Editor), Thegeology of North America, vol. K-2. Quaternary nonglacial geology;conterminous U.S. Geological Society of America, Boulder, CO, pp. 629-650.

Cronin, T.M., 1991. Pliocene shallow water paleoceanography of the North AtlanticOcean based on marine ostracodes. Quaternary Science Reviews, 10: 175-188.

Daniels, R.B., Buol, S.W., Kleiss, H.J. and Ditzler, C.A., 1999. Soil Systems in NorthCarolina, Technical Bulletin 314. Department of Soil Science, North CarolinaState University, Raleigh, N. C., 118 pp.

Dowsett, H.J. and Cronin, T.M., 1990. High eustatic sea level during the middlePliocene; evidence from the Southeastern U.S. Atlantic Coastal Plain; with Suppl.Data 90-13. Geology, 18(5): 435-438.

Dowsett, H.J. and Poore, R.Z., 1991. Pliocene sea surface temperatures of the NorthAtlantic Ocean at 3.0 Ma. Quaternary Science Reviews, 10: 189-204.

Gohn, G.S., 1988. Late Mesozoic and early Cenozoic geology of the Atlantic CoastalPlain: North Carolina to Florida. In: J.A. Grow (Editor), The Atlantic ContinentalMargin: U. S. The Geological Society of America, Denver, pp. 107-130.

Harris, W.B. and Zullo, V.A., 1991. Eocene and Oligocene Stratigraphy of the OuterCoastal Plain. In: J.W. Horton, Jr. and V.A. Zullo (Editors), The geology of theCarolinas; Carolina Geological Society fiftieth anniversary volume. University ofTennessee Press, Knoxville TN, pp. 251-273.

McCachren, C.M., 1978. Soil Survey of Robeson County, North Carolina. USDA-SoilConservation Service, North Carolina Agricultural Experiment Station andRobeson County Board of Commissioners, U. S. Government Printing Office,Washington, D.C., 68 pp.

Mitrovica, J.X., 2003. Recent controversies in predicting post-glacial sea-level change.Quaternary Science Reviews, 22: 127-133.

North Carolina Geologic Survey, 1991. Detailed geologic map of North Carolina andmap legend. North Carolina Geological Survey, Raleigh NC.Gis.enr.state.nc.us/sid/bin/

21

Owens, J.P., 1988. Geology and Tectonic History of the Lower Cape Fear River Valley,Southeastern North Carolina. U. S. Geological Survey Professional Paper 1466-A.United States Government Printing Office, Washington, D.C., 60 pp.

Owens, J.P., 1989. Geologic map of the Cape Fear region, Florence 1° X 2° Quadrangleand northern half of the Georgetown 1° X 2° Quadrangle, North Carolina andSouth Carolina, Map I-1948-A. United States Geological Survey, Reston, VA.

Owens, J.P. and Sohl, N.F., 1989. Campanian and Maastrichtian depositional systems ofthe Black Creek Group of the Carolinas. Carolina Geological Society Field TripGuidebook. North Carolina Geological Survey, Raleigh, N.C.

Peltier, W.R., 2002. On eustatic sea level history: Last Glacial Maximum to Holocene.Quaternary Science Reviews, 21(1-3): 377-396.

Rename, J., Faure-Muret, A. and Odin, G.S., 2001. International Stratigraphic Chart,Second Edition. International Union of Geological Sciences.

Riggs, S.R. and Belknap, D.F., 1988. Upper Cenozoic processes and environments ofcontinental margin sedimentation: eastern United States. In: R.E. Sheridan andJ.A. Grow (Editors), The Atlantic Continental Margin: U. S. The GeologicalSociety of America, Denver, pp. 131-176.

Snyder, S.W., Snyder, S.W., Riggs, S.R. and Hine, A.C., 1991. Sequence stratigraphy ofMiocene deposits, North Carolina continental margin. In: J.W. Horton, Jr. andV.A. Zullo (Editors), The geology of the Carolinas; Carolina Geological Societyfiftieth anniversary volume. University of Tennessee Press, Knoxville, TN, pp.263-273.

Sohl, N.F. and Owens, J.P., 1991. Cretaceous stratigraphy of the Carolina Coastal Plain.In: J.W. Horton, Jr. and V.A. Zullo (Editors), The Geology of the Carolinas. TheUniversity of Tennessee Press, Knoxville, pp. 191-220.

Soller, D.R., 1988. Geology and tectonic history of the lower Cape Fear River Valley,southeastern North Carolina; Surface and shallow subsurface geologic studies ofthe Carolina coastal plains. U.S. Geological Survey Professional Paper 1466-A.U.S. Government Printing Office, Washington, DC, 60 pp.

Soller, D.R. and Mills, H.H., 1991. Surficial geology and geomorphology. In: J.W.Horton, Jr. and V.A. Zullo (Editors), The Geology of the Carolinas. TheUniversity of Tennessee Press, Knoxville, TN, pp. 290-308.

Walker, H.J. and Coleman, J.M., 1987. Atlantic and Gulf Coastal Province. In: W.L. Graf(Editor), Geomorphic Systems of North America, Centennial Special Volume 2.Geological Society of America, Boulder, CO, pp. 51-110.

22

Ward, L., W., Bailey, R., H. and Carter, J., G., 1991. Pliocene and early Pleistocenestratigraphy, depositional history, and molluscan paleobiogeography of the coastalplain. In: J.W. Horton, Jr. and V.A. Zullo (Editors), The Geology of the Carolinas.The University of Tennessee Press, Knoxville, TN, pp. 274-289.

Ward, L.W. and Blackwelder, B.W., 1980. Stratigraphic revision of upper Miocene andlower Pliocene beds of the Chesapeake Group, middle Atlantic Coastal Plain,Report: B 1482-D. U. S. Geological Survey Bulletin. U. S. Geological Survey,Reston, VA, D1-D61 pp.

Yokoyama, Y., Fifield, L.K., Lambeck, K., De Deckker, P. and Johnston, P., 2000.Timing of the Last Glacial Maximum from observed sea-level minima. Nature,406(6797): 713-716.

23

Figure 1. Generalized model of the delta to shelf lithofacies developed by Sohl andOwens (1991) for the Cretaceous of the Carolinas.

24

Figure 2A. Juniper Bay area as shown in the map of Owens (1989). Juniper Bay locationis shown by the blue oval. See Table 2A for key to geologic units.

Dillon County,South Carolina

Robeson County,North Carolina

Bladen Co.

Robeson Co.

25

26

(Sohl and Owens, 1991.)

27

0 2 , 0 0 0 4 , 0 0 01 , 0 0 0 M

Figure 4. DOQQs of the area around Juniper Bay showing the NE-SW trending featuressuggesting old coastal dune systems.

US-74

HogSwamp

Big IndianSwamp

28

Figure 5. Digital Elevation Model of the setting of Juniper Bay (data fromhttp://www.precisionag.ncsu.edu/data/usgs/dem/nad83m/robeson.zip, 10Xexaggeration).

29

Figure 6. Locations of the 29 cores collected at Juniper Bay in August 2000. Bay rimindicated approximately by the oval line. Cores discussed in the text are marked as JI01,JI02, JI03, JI06, JI08, and JI09; other cores are indicated just by number.

30

Figure 7. Core stratigraphy. S = sand; LS = Loamy sand; C = Clay; CL = Clay loam;SCL = Sandy clay loam; VCSCL = Very coarse sandy clay loam; SL = Sandy loam;CoSL = Coarse sandy loam; SC = Sandy clay; SiCL = Silty clay loam.

JI010

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

Dep

th (m

)JI02

0

1

2

3

4

5

6

JI030

1

2

3

4

5

6

JI060

1

2

3

4

5

6

JI080

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

JI090

1

2

3

4

5

6

7

8S LS

CL C

VCSCL SCL

SL CoSL

SC SiCL

31

Figure 8. Elevation of the first clay layer inside (A) and outside (B) the bay.

32

Figure 9. Coarse sand in six cores inside the bay compared to Core 28 outside the bay.

33

Figure 10A. Clay content of nine cores inside Juniper Bay.

Figure 10B. Clay content of five cores outside Juniper Bay.

34

Table 1. Portion of the geologic time scale important to Juniper Bay geology (fromRename et al., 2001).

Geologic Time Scale SubdivisionsEon Era Period Epoch

ESTIMATEDAGE (YRS X 106)

Holocene 0-0.01QuaternaryPleistocene 0.01-2Pliocene 2-5Miocene 5-24Oligocene 24-38Eocene 34-55

CenozoicTertiary

Paleocene 55-66Maastrichtian 66-71Campanian 71-84Santonian 84-86Coniacian 86-89

Late

Turonian,Cenomanian

89-99

Cretaceous

Early 99-142Jurassic 138-205

Phanerozoic

Mesozoic

Triassic 205-250

35

Table 2A. Key to Figure 2A. North Carolina Geologic Survey: general geology of NorthCarolina near Juniper Bay and the reference bays (North Carolina Geological Survey,1991).Key to GeologicUnits

Period Epoch Formation

Tpyw Tertiary Plio-Pleistocene WaccamawTpy Tertiary Pliocene Yorktown and Duplin,

undivided (Duplin south of theNeuse River)

Tp Tertiary Upper Miocene PinehurstKp Cretaceous Upper Cretaceous-

MaastrichtianPeeDee

Kb Cretaceous Upper Cretaceous,Maastrichtian-Campanian

Black Creek

Km Cretaceous Upper Cretaceous, Santonian MiddendorfKc Cretaceous Upper Cretaceous, Santonian Cape Fear

Table 2B. Key to Figure 2B; Owens (1989) map of the detailed geology of the Cape FearRegion, scan of the portion showing mapping units around Juniper Bay.Key to Geologic

UnitsPeriod Epoch Formation

Qwa Quaternary Upper Pleistocene WandoQs Quaternary Upper Pleistocene SocasteeQph Quaternary Lower Pleistocene PenholowayQw Quaternary Lower Pleistocene WaccamawTb Tertiary Upper Pliocene Bear BluffTd Tertiary Lower Pliocene Duplin

36

Table 3. Stratigraphy of southeastern North Carolina, with geologic age of each unit, unitname, description, and environment of deposition. Developed from information in Owens(1988, 1989) and Sohl and Owens (1991), and Dowsett and Poore (1991).Formation Age Description Depositional environment

Contemporary flood plain Holocene, <10,000ybp

Active flood plain Fluvial, no Carolina Bays

Wando Formation Upper Pleistocene,~90,000 ybp

Terrace Fluvial, no Carolina Bays

Socastee Formation Upper Pleistocene,~200,000 ybp

Terrace Fluvial, dune fields, CarolinaBays

Penholoway Formation (fluvial facies inthe Cape Fear River Valley—also in theuplands below 21 m elevation, 40 kmSE of Juniper Bay)

Lower Pleistocene,>760,000 ybp

Terrace, Sand--15 m thick Fluvial, dune fields, CarolinaBays

Waccamaw Formation Lower Pleistocene,~1,750,000 ybp

Terrace, sand with basalgravels; ~14 m thick. Locally,Mulinia fossil beds

Fluvial, dune fields, CarolinaBays

Bear Bluff Formation Very late Pliocene-very earlyPleistocene, 2,40,000-1,800,000 ybp

Terrace, sands and calcareoussilts, overlies upper Cretaceousunits, aragonite mollusks

Fluvial, shallow marine,poorly preserved CarolinaBays

Yorktown Formation/Duplin Formation Early Pliocene,3,000,000-3,400,000ybp

Shelly, medium- to coarse-grained sand, sandy marl,bluish gray. Up dip,interbedded thin gray clay andsilt and yellow sand. Lack ofweathering in upper part of theDuplin indicates extensiveerosion. Some sands arethicker, gravelly, cross-bedded.Some are burrowed. Oftenfossiliferous; fossil beds are upto 3 m thick near Lumberton,fills channels cut intounderlying Cretaceousformations.

Complex marginal marineenvironment withinterfingered marine andnon-marine sands. AbundantCarolina Bays.

PeeDee Formation Upper Cretaceous-Maastrichtian

Massive. Dark greenish-gray togray fine sand, sparinglymicaceous, glauconitic, withfew marine clay beds.Calcareous. Fragmented shells.Common burrows. Irregularsurface due to post-depositionalerosion.

Shelf

Donoho Creek Formation Upper Cretaceous-Lower Maastrichtian

More massively bedded ingeneral, black clay andmicaceous sand, fossils deeperin the section, glauconitic.Abraded shells. Reworkedsands, bones, and teeth atbottom.

Delta front-prodelta to shelfup section. Very erodedsurface. Much has beenstripped from the Cape Feararch area.

Bladen Formation Late Campanian Thin laminated black clays withlaminae of light colored verymicaceous sands; with incisedchannels. No marine fossils,lignitized wood fragments.

Shelf (deeper water than theTar Heel)

BlackCreekGroup

Tar Heel Formation Lower Campanian Thin black carbonaceous bedsof clay with cross-bedded lightcolored micaceous sand;pyritized wood fragments;marine bones and sharks’ teeth,no bivalve fossils

In Cape Fear River area, deltafront with minor marineinfluence

Middendorf Cretaceous-Santonian Interbedded black clay andsand

Delta front

Cape Fear Cretaceous-Coniacian Sand and clay beds of varyingthickness, scattered mica

Upper delta plain

37

Table 4. Core observations inside the bay. JI01, 02, 03, 06, 08, and 09 stratigraphy is shown inFigure 7. The lower part of the table shows details notes in additional cores.

Cores fromFig 7

Depth(m)

Observations and measurements

0-5.8 Sandy textures with one thin (9 cm) SCL layer; rhythmites at 3.4m

>5.8 Grey/black colors; silty clay 5.8-7.4 ; coarser sand below 7.4 m

JI01

6.7-8.8 Rhythmites: fines separated by thin sand lensesJI02 0-5.8 Sandy textures with three (10, 40, 20 cm) SCL layers; Buried A

at 1.1 m0-5.8 Sandy textures with Clay/SCL between 1.2-3.7 mJI033.7 Wood fragments

JI06 0-5.8 Buried A at 1.9; wood fragments at 1.9, 2.5, 4.3, 5.5 m; sand 0-2.1 and 5.1-5.8; Clayey textures 2.1-5.1

0-7.1 0-0.7 sandy, 0.7-6.7 clayey textures; 6.7-7.1 sandy textures3.8-5.9 Granular structure, shell ghosts, burrows, root channels5.9-6.6 Rhythmites (sand lenses)7.1 Pebbles, 25% coarse sand7.1-14.9 Rhythmites; clayey 7.1-9.1, 11-13; sandy 9.1-11, 13.1-14.911.9-13.4 Abundant mica

JI08

>9.1 Coarse sand0-7.3 0-3 m sandy; 3-4.3 clayey; 4.3-6.1 sandy; 6.1-7.3 clayey2.0 Buried A3.8-5.3 Shells (Mulinia, Chesapecten), shell ghosts3.8-7.3 Reaction to HCl

JI09

11-13 Abundant micaOther cores

0-7.4 Sandy to 2.8 m with 20 cm of SCL at 1.4 m, SC 2.3-2.6 m 2.8-7.4 clayey textures with sandier textures at very bottom.

JI04

5.8 RhythmitesJI05 0-5.8 Buried A at 2.5 m; wood fragments at 1.8, 2.5, 4.1, 5.5 m. Sandy

with clayey textures 0.9-2.1 m and 2.3-3 m.0-5.8 Sandy to 2.8 m; clayey textures to 5.7 with sand at very bottom;

Buried A at 0.7, 2.4, and 3.4 mJI07

3.4 –5 Clay with vertical seams of sand0-5.8 Sandy textures with SiCL between 0.75 and 1 m and clayey

textures 2.5-4 m; very coarse sand at 4.3 mJI10

5.5 Rhythmites (sand lenses)JI11 0-5.8 Buried A at 1.2 m; shell ghosts at 3.3 m; sandy textures with

loam: 30 cm at 1.7 and 40 cm at 2.6 (with root fragments).0-7.3 Sandy with clayey textures 1.2-2.5 m and 6.6-7.3 m> 6.6 Black clayey textures

JI13

4-5.9 Rhythmites

38

Chapter 3

HISTORICAL RECORDS OF AGRICULTURAL PRACTICES THAT AFFECTEDHYDROLOGY AND SOILS AT JUNIPER BAY

J.M. Ewing, M.J. Vepraskas, and C.W. ZannerINTRODUCTION

Wetland restoration is being conducted on an increasing scale throughout the U.S.In the southeastern U.S., wetlands that were cleared and drained for agriculture in the1930’s or later, are now having their ditches plugged to restore the original hydrologyand wetland trees are also being planted to restore the original vegetation. Suchrestoration is expensive and any information that increases the likelihood of success is ofvalue. We are involved in a current restoration of a Carolina Bay wetland in NorthCarolina. As part of this effort we assembled a record of historical data on land use thatprovided additional insight into practices that have affected both the soils and hydrology.

In our experience, historical data on land use is frequently not used in planningrestoration activities. Water sources such as surface inflows and outflows can affectrestoration and should be know before hydrologic alterations are designed. Existence ofperennial natural lakes suggest groundwater inflow is an important contribution to awetland’s hydrology. However, small lakes or ponds may not appear on all soil maps,especially if the area has been drained for agriculture. Many construction or demolitionprojects, land clearing, mining and farming can leave the landscape looking unaltered,but examinations below the surface can show dramatic soil variability due to humanactivities. For example, soils developed from mine spoils have been shown to be morevariable between 1 to 10 m depths than natural soil (Shafer, 1979). In soil researchstudies it is important to have a grasp on the variability of the soil so that the correctnumber of samples can be taken. In areas known to be altered by human activity, it canbe beneficial to collect historical land use data to document where and when alterationsoccurred that could explain variation.

The North Carolina Department of Transportation (NCDOT) bought 750 ac (256ha) of drained agricultural land in 1999, near Lumberton, North Carolina. This land waspurchased with the intent of returning it to its natural state so that the NCDOT would gainwetland mitigation credits. The NCDOT awarded a seven-year grant to North CarolinaState University, Soil Science Department, in 2000, to evaluate and help developmethods of restoration that will insure success. The Carolina Bay being restored is calledJuniper Bay. Since one of the hopeful outcomes of this project is the restoration of soilstypically found in Carolina Bays we had to describe the soils as they were after drainageand agriculture production and to examine soils of undrained Carolina bay soils. ThisCarolina bay wetland was altered thought ditching and other to allow agricultureproduction to occur. However former land owners also reported other activities that hadoccurred that may have altered soil properties.

39

The purpose of this study was to assemble an historical record of land usespanning the last 80 years to find potential sites of human induced variability and discussimplications on wetland restoration.

MATERIALS AND METHODS

Site Description

Carolina bays are elliptical depressions in the landscape that are orientated alongthe long axis SE to NW (Prouty, 1952). They range in size from 40 ft (10 m) to>2.5miles (4 km) along the long axis. The bays are usually surrounded by a sandy rimand have a high amount of organic matter within the depression (Johnson, 1942). Theextent of these bays range from Northern Florida to Delaware with the highestconcentration in North and South Carolina. Estimates on the number of these bays are ashigh as 500,000 (Johnson, 1942), but the actual number maybe less than 100,000(Nifong, 1998). Many theories for bay formation have been proposed, from the mostpopular, meteor impact (Johnson, 1936), artesian springs (Prouty, 1952; LeGrand, 1952),whale wallows (Grant 1945), and ice flows (Bliley and Burney, 1988). Currently themost plausible explanation is that originally there was a slight depression in the landscapewith a shallow aquitard that allowed the water table to be held above the surface.Prevailing winds then shaped the depression into the now familiar orientated shape(Thom, 1970; Odum, 1952). During the past century agricultural and communitydevelopment have led to the drainage and use of these bays. It is estimated that 50% ofall Carolina bays were drained and developed in some manner in Bladen County, NC by1982 (Weakley and Scott, 1982). This figure would be higher if other managementpractices such as logging were included. As these bays are used for agriculture and otheractivities, their defining characteristics of sand rims and organic surfaces, become blurredinto the surrounding landscape.

The Carolina bay in question is called Juniper Bay and is located approximately 7miles (10km) south of Lumberton, North Carolina. The soil survey of Robeson County(Fig.1), shows Juniper Bay with areas of Ponzer (Loamy, mixed, dysic, thermic TerricHaplosaprists), Leon (Sandy siliceous, thermic Aeric Alaquods), Pantego (Fine-loamy,siliceous, semiactive, thermic Umbric Paleaquults), Rutlege (Sandy, siliceous, thermic,Typic Humaquepts) (McCachren, 1978). An earlier soil survey (Hearn, 1909) identifiedthe soils in Juniper Bay as Portsmouth fine sandy loam.

Sources of Information

There were several sources of information that we used to conduct the historicalland use search. They included aerial photos from the United States Aerial PhotographyService Office, the former landowner, Mr. Robert Freeman, courthouse documents andthe National Railway Historical Society (NRHS). Aerial photos were available from1938, 1952, 1961, 1966, 1972, 1981, 1991, and 1997, and were obtained from the USDA-FSA Aerial Photography Field Office, 2222 W 2300 S, Salt Lake City, UT 84119-7619.Mr. Freeman provided a wealth of information about Juniper Bay before it was drained.

40

He was also the person who drained and placed Juniper Bay into agricultural production.Documents found from the Robeson County courthouse provided a history of ownershipand use. The NRHS provided information concerning a railroad that ran through Juniperbay and how the tracks were constructed.

RESULTS AND DISCUSSION

1911-1938: Railroad Years

A Robeson County road map found at the county court house, dated 1922,showed a creek flowing out of Juniper Bay that was initially unnoticed in the aerialphotos. This creek flowed out of the NE corner and was originally the head water ofLittle Indian swamp which flowed south into Big Indian Swamp which fed into Ashpoleriver and finally into the Lumber River which is an important river in the area. A faintoutline of this creek was discovered in a few of the older aerial photos upon closeexamination. The 1938 aerial photo (Fig. 2) shows that Juniper Bay was heavilyforested. The vegetation supported abundant wildlife. Mr. Freeman told of seeing deer,black bear, beavers, bobcats, snakes, foxes, otters, and hundreds of geese and ducks whenhe hunted there before draining Juniper Bay. The densely forested bay contained largeAtlantic White Cedars, Bald Cyprus, Tupelos, and various Bay trees.

While virtually undeveloped, the site did have a railroad line running through it ina north-south direction. Mr. Freeman remembered stories about how his father wouldhop the train in Fairmont and ride it10 miles into Lumberton to conduct business. A mapof the Seaboard Airline Railway (SAL RY) in 1911 (Prince, 1966) showed a routerunning from Lumberton through Ashpole (Fairmount) and to Pee Dee (just NE ofMarion S.C.). This route would run through Juniper Bay and is identified (Fig. 3) in the1909 soil survey of Robeson County (Hearn, 1909). The company that built and ownedthe railroad that ran between Lumberton and Marion, South Carolina, was The Raleigh-Charleston Railroad Company. The Raleigh & Charleston Railroad (R&C) was formedin December of 1905 (Carriker, 1985a). SAL RY bought controlling stock of the R&C inNovember 1911 (Cariker, 1985b), and in 1933, 23 miles of its line between Lumberton,N. C. and Lakeview, S.C. in 1933, and the remainder into Marion in 1946 (Prince, 1966).

The railroad bed was 4 to 5 feet (~1.5m) above the surrounding terrain. Mr.Freeman recalled that there was one ditch that ran along the track, and that it appeared therailroad bed was mostly fill material from the ditch. With the assistance from the NRHS,literature was found that supported the Mr. Freeman’s description of the construction ofthe railroad bed in Juniper Bay. During the time of R & C, the common constructionpractice was to dig a ditch and use the spoil for the railroad bed (Kirkman, 1904). Theditch supplied fill for the bed, and was not dug for drainage or long-term bed stability.Clay was brought to the site to stabilize sections of the bed at a later date. We estimatethat the railroad bed in Juniper bay was elevated to match the surrounding areas, resultingin a bed elevation of 3 to 5 ft (~1-1.5m) above the surface inside the bay. This bed wasalso estimated to have been approximately 10 ft (~3m) wide with an adjacent ditch onone or both sides of the bed. Ditch depth probably ranged from 5 to 10 ft (~2-3 m), but

41

an exact depth could not be determined. It is still possible to see where the location ofthe old railroad in current aerial photos. In addition, a slight crest across the bay wherethe bed used to be can still be seen when in the field. We have also found several rustedrailroad spikes along the old railroad transect during our field studies.

While the railroad’s tracks are visible in the 1938 photograph, it was apparentlyabandoned by that time. It is uncertain if the tracks were removed by 1938. During oursoil sampling, pits placed where the old railroad bed was did have evidence ofdisturbance by excavation and filling to a depth of one meter (Fig. 4). This disturbancecould be traced along the railroad line, and was considered to be of small extent.

1938-1966: Forest Harvest

Aerial photos from 1958 showed little change in Juniper Bay since 1938, but by1966 some major changes had occurred. M. Carr Gibson of Canal Industries bought theland from Lawrence Ballard in the mid 1960’s to harvest trees. The 1966 aerial photo(Fig 5) shows an outline through the vegetation where drainage ditches will be located,and also shows that some of the vegetation has been thinned. One ditch runs parallel,possibly adjacent to the old railroad bed. Another ditch will run SW to NE and withanother perpendicular ditch to run SE to NW. Canal Industries harvested timber from theentire the bay except for the area around the center of the bay that contained a shallowlake and parts of this lake can be seen in the 1966 photo.

By the time Canal Industries sold Juniper Bay to Robert Freeman Sr., a drainagesystem on the NW side of Juniper Bay had been constructed (Fig. 1). This drainagesystem included a perimeter drainage ditch surrounding the NW end of the bay, fivelateral drainage ditches running parallel to the railroad bed on the north side of theNW/SE ditch, and one ditch that ran the width of the bay. It is uncertain if the railroadbed had been leveled at this time but the drainage ditch along the old railroad bedprobably still served as a collector to the main outlet. During the time Canal Industriesowned Juniper Bay, the Robeson Co Soil Survey was being conducted. Willie Spruill, asoil surveyor who helped with the survey, has stated that often Carolina bays were sothick with vegetation that it was impossible to conduct a thorough survey through the bayitself. Often the Carolina bays were mapped extrapolating soil map units into theCarolina bays from adjacent lands. Vegetation patterns visible on aerial photos wereused to estimate map unit boundaries within the bays (Soil Survey Manuel, 1993, Buol etal., 1997). Mr. Spruill could not recall if this is how Juniper Bay was mapped and it isuncertain how accessible Juniper Bay was at the time of mapping, however, the 1966 andthe 1971 aerial photographs show that the soil delineations follow vegetation patternsfairly well.

1975-1981: Clearing and Cultivation

Mr. Freeman’s father purchased Juniper bay in October 1975 and drainage of thebay for agriculture began during a dry summer in 1979. This was no small operation, andover 15 individuals were employed to operate 10 bulldozers, 3 trackhoes, and a dragline.

42

Over a period of 9 months in 1979, the railroad bed was leveled and the parallel ditcheswere filled. Mr. Freeman said that when he was ditching the bay, there were manycreosote crossties that remained, but no rails. Lines where the lateral, main and perimeterditches were to be located, evenly spaced, were cut with handheld equipment includingchainsaws, machetes, and bush-axes because the soils were too wet to support heavyequipment. While cutting over 30 miles (48 km) of trails for the ditches, workers wouldclimb up in the trees to escape the stifling heat and mosquitoes during lunch and workbreaks.

Once trails for the ditches were cut, a trencher was used to excavate the ditchesinside the bay. Workers had to lay logs and other debris ahead of the trencher so it had a“solid surface” to work on. A dragline was used to dig the perimeter ditch on the sandyrim of the bay. During the ditching and clearing operations, many preserved trees, logsand roots were pulled from the muck with tree rakes (Fig 6). For years after the clearingoccurred, any time the fields were tilled, wagonloads of roots and debris had to beremoved. During our sampling of soils in 2001 we did find buried roots of trees in manypits.

After enough drainage had occurred to allow the soils to support the weight ofheavy equipment, debris was windrowed with bulldozers into piles 40 to 50 ft wide (~15-20m), that ran the length of the bay (Fig. 7). The debris was then burned, with some pilesburning for 2 to 3 years. The debris fires ignited the peat that lined the bay floor, and insome areas burned the surface was lowered 2 to 3 ft (50-90cm) to either a mineral layeror the water table. This created a depression that filled with water and was inaccessiblewith any equipment (Fig. 8). To drain the depressions, an additional ditch had to be dugor the depression had to be filled. Fill material would come from another ditch, locatednear the depression. This is why the lateral ditches do not match up across the length ofthe bay, the cuts are of uneven size, and the distance between cuts decreased. After theditches were in place and sufficient drainage had occurred, spoil from the ditches wereused to shape all the fields such that the center of the field was approximately 18 inches(45cm) higher than the edges to increase surface runoff. This practice is locally called“crowing” or “turtle backing.”

1981-2000: Agriculture Production

The aerial photo from 1981 (Fig.9) shows almost all of the bay completelydrained and in agricultural use except for a small corner in the NE section. This was andis the wettest and lowest area in the bay. This was re-emphasized when a landowneradjacent to Juniper Bay on the NE cleared and ditched a small parcel of land and tied intoJuniper Bay’s drainage system. Unfortunately for this farmer, his land was lower thanthe main outlet of Juniper Bay, and thus his new field became a semi-permanent pond.That landowner eventually disconnected from Juniper Bay’s drainage system. The NEcorner of Juniper bay was cleared and drained in 1986, and put into production thefollowing year.

43

Currently, 2003, Juniper Bay is drained though one outlet point (Fig.10). There isa perimeter ditch around the whole bay, and two main ditches that run SW/NE and onethat runs SE/NW, which are approximately 9 feet (3m) deep and 20 feet (6m) wide.Perpendicular to the SW/NE main ditches are lateral drainage ditches that areapproximately 3 ft (1m) deep and 5ft (1.5m) across. Mr. Freeman stated that ditcheswere maintained as needed yearly. Every 5 or 6 years all ditches were cleaned using apiece of equipment, called a Dondi ditcher (Fig. 11), that dredged out the ditch and slungthe spoil out over the field, where it was disked in.

Mr. Freeman maintained excellent records and has soil test results dating back to1976 giving us an idea of what the chemical variability was. The 1976 soil test analysisof the plow layer from various (unknown) locations across the bay, conducted byBrookside Farms Laboratory Association, Inc., showed that pH ranged between 3.6 and4.8, cation exchange capacity 2.76-14.04 meq 100g-1, 2.7 to 11.5 % organic matter, 145 –2240 lb/ac P, 279-698 lb/ac Ca, 26-363 lb/ac K, and 31-51 % base saturation. Soil testreports from 1978 and 1979, by A & L Eastern Agricultural Laboratories, Inc., ofRichmond, Virginia, show approximately the same values. Mr. Freeman estimated that15 to 20 cumulative tons per acre of lime have been applied over the years in addition tothe yearly-recommended fertilizer rate. Crops grown in Juniper Bay included; soybeans,corn, tobacco, cotton, oats, millet, wheat, and vegetable crops like lettuce and okra. Oneyear, a leasing farmer raised winter wheat and burned the stubble off leaving a blanket ofash that looked like snow. Of course this re-ignited some of the peat, burning some areasdown to the water table, most notably in the NE corner.

Use of Historical Data

Historical information has provided insight into spatial heterogeneity of the soilsand hydrology that we have at Juniper Bay. Successful restoration depends on areturning the site’s hydrology to what it was prior to ditching. The historical record hasgiven us a glimpse of the past hydrology. The 1922 Robeson, County map documentedthe creek on the NE side of the bay that was historically a source of surface waterremoval. The aerial photos and personal accounts indicated the existence of a naturallake in the center of the organic soils. This was unexpected because none of theneighboring bays have lakes or even organic soils. This suggests that Juniper Bay wasreceiving subsurface inputs of water, unlike the neighboring bays, which may have beenrain-fed. Our hydrologic measurements are confirming that groundwater upflow(discharge) does appear to be occurring on the north side of the organic soils.

The historical record has also shown that human activities have affected soilvariability in Juniper Bay. We were able to document processes that affect organic soilsubsidence, and estimate that 2 to 3 ft (50-90cm) of subsidence has occurred. It is wellknown that organic soils subside following drainage (Everett, 1983). Processesresponsible for this include a rapid settling of material as it loses the buoyant forceprovided by water after drainage (primary subsidence), oxidation of organic matter, andshrinkage of the organic material on drying (both termed secondary subsidence). Thehistorical record shows that the primary subsidence can be increased by compaction from

44

the heavy equipment used to clear the land. More importantly, fire increases oxidationrate of organic matter. The historical record has shown that fire events have occurred inJuniper Bay more than once and one fire burned for 2 years. We found evidence of fire,including charcoal and ash, in several soil profiles throughout Juniper Bay. We havedated charcoal and other buried vegetation to 8400 and 3100 years before present(Zanner, 2001 personal communication). The burning also explains why some of thewettest areas in the bay were devoid of organic soils. Estimating the amount ofsubsidence that has occurred is useful to predict the elevation of the original water table,and the potential elevation after restoration.

Construction of the railroad and the current and previous ditching systems hasaltered the soils to depths of 1 m and more. Historical aerial photos providedapproximate dates showed where the ditching occurred. We were able show how the fillfrom the railroad bed changed the soil profile, which allowed us to factor in similar areasduring studies of the soils and hydrology. The present soil material at the surface is notnecessarily a product of the original soils owing to the spreading of dredge materialduring canal construction. The surface of Juniper Bay has been shaped several timessince drainage to promote surface drainage, and ditch maintenance brought subsurfacematerial to the surface and broadcast it over the field.

Our historical records also included the dating of agricultural developmentincluding application of lime and fertilizer. Leaching of agricultural chemicals, notablyphosphorus, to the groundwater is currently of great interest because it reduces quality ofsurface waters. The historical record provided specific dates for when chemical additionsbegan. This information is being used to evaluate the rate and depth of chemicalmovement through the soil profile over 15, 20, and 30 years of agricultural additionsacross the bay.

By establishing the presence of spatial variability prior to drainage evaluation ofrestoration efforts should not expect uniform results over the entire area. Although manyearly soil surveys identified wetland areas only as swamps, marshes, etc., important soilproperties were not identified or spatially mapped. Many of the soil properties appearrelated to underlying hydrology that will again influence restoration results.

45

REFERENCES

Bliley, D.J. and D.A. Burney. 1988. Late Pleistocene climatic factors in the genesis of aCarolina Bay. Southeastern Geo. 29:83-101.

Buol, S.W., F.D. Hole, R.J. McCracken, and R.J. Southhard. 1997. Nature of soil cover:polypedons, soilscapes, and mapping units. In Soil Genesis and Classification. 4th

ed. Iowa State University Press, Ames IA.

Carriker, S.D. 1985a. Railroading in the Carolina Sandhills: Vol 1: The 19th Century(1825-1900). Heritage Publishing Company, Matthews, N.C.

Carriker, S.D. 1985b. Railroading in the Carolina Sandhills: Vol 2: The 20th Century(1900-1985). Heritage Publishing Company, Matthews, N.C.

Everett, K.R. 1983. Histosols. In Pedogensis and Soil Taxonomy II. The Soil Orders.L.P. Wilding, N.E. Smeck and G.F. Hall eds. Elsevier, Amsterdam. pp. 1-53.

Freeman Jr, R. 2001, Personal Interview.

Grant, C. 1945. A Biological explanation of the Carolina Bays. Sci. Monthly 61:443-450.

Hearn, W.E. 1909. Soil Survey of Robeson County, North Carolina. USDA Bureau ofSoil. Govt. Printing Office.

Johnson. D.W. 1936. Origin of the Supposed Meteorite Scares of the Carolina. Science.84:15-18.

Johnson, D.W. 1942. Origin of the Carolina Bays. Columbia University Press.Kirkman, M.M. 1904. Building and Repairing Railways: Supplement to TheScience of Railways. The World Railway Publishing Company, NewYork.pp.155.

LeGrand, H. E. 1953. Streamlining of the Carolina Bays. J. Geol. 61:263-274.

McCachren, C.M. 1978. Soil Survey of Robeson County, North Carolina. USDA-SCS.National Archives Aerial Photography, 1938. Aerial Photo Robeson County, NC.ACT-53-3630.

Nifong, T.D. 1998. An ecosystem analysis of Carolina Bays in the Coastal Plain of theCarolinas. Ph.D. Dissertation. University of North Carolina, Chapel Hill, NorthCarolina.

46

Odum, H. T. 1952. The Carolina Bays and a Pleistocene weather map. Am. J. Sci.250:263-270.

Prince, Richard. 1966. Seaboard Air Line Railway; Steam Boats, Locomotives andHistory.Indiana University Press, Bloomington, IN. pp.121.

Prouty, W.F. 1952. Carolina Bays and their Origin. Bulletin of the Geological Societyof America. 63:167-224.

Schafer, W.M. 1979. Variability of mine sites and natural soils in southeastern Montana. Soil Sci. Soc. Am. J., 43:1207-1212.

Soil Survey Division Staff, 1993. Soil Survey Manual. United States Department ofAgriculture, Washington D.C.

Thom, B.G. 1970. Carolina Bays in Horry and Marion Counties South Carolina. Geol.Soc.Amer. Bull. 81:783-813.

USDA-FSA Aerial Photography, 1988. Aerial Photo Robeson County, NC (10-23-88)USDA 40 1188-74.

USDA-FSA Aerial Photography, 1981. Aerial Photo Robeson County, NC (1-5-81)USDA 40 179-165.

USDA-FSA Aerial Photography, 1972. Aerial Photo Robeson County, NC (3-9-72)A 20 172-212.

USDA-FSA Aerial Photography, 1966. Aerial Photo Robeson County, NC (3-9-66)ACT-2GG-16.

Weakley, A.S. and S.K. Scott. 1982. Natural features summary and preserve design forCarolina Bays in Bladen and Cumberland Counties, NC. Unpublished report toNC Natural Heritage Program, The Nature Conservancy, and Earthlines, Inc.,Raleigh,NC.

Zanner, C.W. 2001. Personal communication.

47

+

Figure 1. Aerial photo 1972 and Soil Survey map of Juniper Bay, Sheet 61 (USDA-

SCS,1978).

LegendCo- Coxville loamDn- Dunbar sandy loamLaB- Lakeland sand 0-6%Le- Leon sandLy- Lynchburg sandy loamGoA- Goldsboro loamy sand, 0-2%NoA- Norfolk sandy loam, 0-2%NoB- Norfolk sandy loam, 2-6%Pg- Pantego fine sandy loamPm- Plummer and OsierPoB- Pocalla loamy sand, 0-3%Pr- Ponzer muckRa- Rains sandy loamRu- Rutledge loamy sandTo- Torhunta loamWaB- Wagram loamy sand, 0-6%WkB- Wakulla sand, 0-6%

48

Figure 2. Aerial photo of Juniper Bay in 1938 (National Archives and Records

Administration).

49

Figure 3. The 1909 Soil Survey of Robeson County (Hearn, 1909) shows the Raleigh &

Charleston Railroad going through Juniper Bay. The “R.R.” on the map is located inside

the boundaries of Juniper Bay.

Legend

Nsl- Norfolk fine sandy loamNfs- Norfolk fine sandPl- Portsmouth fine sandy loamS- Swamp

50

Figure 4. Soil profile showing a prior ditch and subsequent fill material from the railroad

transect. Arrow points to the interface between undisturbed soil and ditch fill. Scale is in

cm.

51

Figure 5. Aerial photo from 1966(USDA-FSA).

52

Figure 6. Log-rake pulling log from the soil in the Blacklands of North Carolina (with

permission S.W. Buol).

53

Figure 7. Debris from clearing piled into windrows in the Blacklands of North Carolina

(with permission S.W. Buol).

a. b

Figure 8. a). Windrows burning down to mineral. The white areas are ashes from the

most intense fire and the black areas to the right are areas of largely unburned organic

soil. b). Loss of approximately 30 miles (48km) organic soil after a wild fire in the

Blacklands of North Carolina (with permission S.W. Buol).

.

54

Figure 9. Aerial photo of Juniper Bay from 1981(USDA-FSA).

Figure 10. Aerial photo of Juniper Bay from 1993 (USDA-FSA) Arrow indicates

drainage outlet.

55

.

Figure 11. Ditch maintenance with a Dondi ditcher.

56

Chapter 4

VEGETATION OF THREE REFERENCE CAROLINA BAYS IN RELATIIONTO SOILS AND HYDROLOGY

B. Lees, J. M. Stucky, M. J. Vepraskas,and T.R. Wentworth

INTRODUCTION

Carolina Bays are unique ovoid depressions distributed throughout the AtlanticCoastal Plain from Florida to Virginia. An estimated 13,000 Carolina Bays exist in NorthCarolina and South Carolina alone. Unaltered Carolina Bays are consistently elliptical,oriented along a northwest to southeast axis, exhibit a sand rim along the eastern side ofthe depression, range in size from less than 50m to 8km (5 miles) across, and containdeepwater or depressional wetland habitat (Sharitz and Gibbons, 1982). Prouty's (1952)estimate of 400,000 - 500,000 total Carolina Bays is now considered a gross overestimateand is thought to have led to a low conservation priority and, consequently, a highexploitation rate (Nifong, 1998). Humans have cleared the native vegetation from andaltered the hydrology in 79% of the Carolina Bays that have been identified (Nifong1998). Collectively, unaltered bays function as wildlife habitat for several endangeredanimals (Clark et al., 1985; Zevelof, 1983), provide habitat for several rare plant species(Weakley, 1982), support an array of unique plant communities (Weakley, 1982), providestormwater storage on a landscape level, and act as carbon sinks (Richardson et al., 1981;Bridgham et al., 1991). Therefore, it is important to understand the structural attributes(vegetation), physical settings (geomorphology), hydrologic regimes, and soil chemistrythat compose the Carolina Bay ecosystem of this area so that we can preserve, manage,and restore these bays to their maximum functional capacities.

There is relatively little known about the relationship between soils, hydrology,and distribution of plant community types in Carolina Bays (Sharitz and Gibbons, 1982;Nifong, 1998; Reese and Moorhead, 1996). Although other peatland wetland systems,such as fens and mires, have been studied extensively, the hydrology, substrate, andvegetation in these ecosystems tend to be distinctly more homogenous than in CarolinaBays (Schalles and Shure, 1989; Malmer, 1986; Vitt and Chee, 1990; Bridgham andRichardson, 1993). Variation in soil chemistry and water table regime in Carolina Baysis associated with an array of distinct plant communities that are typically found inconcentric patterns between the rim and center of individual bays (Reese and Moorhead,1996).

The North Carolina Department of Transportation (DOT) currently is restoring adepressional wetland in a 750 acre Carolina Bay located in Robeson County, NC. ThisCarolina Bay was ditched, drained, and converted to agricultural land in the early 1970s.The success of the restoration project will be determined in the future by comparingcharacteristics of the restored ecosystem at the mitigation site to those of reference sites.While previous research presented general descriptions of Carolina Bay vegetationcommunities and their distribution patterns, the information they provided was not

57

applicable reference data due to the absence of quantitative and comparablemeasurements necessary for determining the success of vegetation community restorationon a given soil type. Kologiski (1977), in a thorough analysis of the Green SwampNatural Area in Brunswick County, North Carolina, described several peatlandcommunities. His research built on previous descriptive studies by including aquantitative analysis of the vegetation communities correlated with soil series. However,his quantifications were used in ordination techniques in order to delineate communitytypes based on soil depth, hydroperiod, and frequency of fire. The quantifications werenot included in his publication. Therefore we were not given the necessary structural orcompositional quantities necessary for measuring the establishment of a specifiedcommunity type at a restoration site, for which our study is intended. In addition, ourresearch focuses on Carolina Bay systems, and Kologiski performed his study in a coastalplain wetland complex. Although inferences can be made regarding between the twosystems, our reference data is from Carolina Bays and suitable therefore suitable for thepurposes of Carolina Bay ecosystem restoration. Schafale and Weakley (1990 and 1991)classified and described pocosin vegetation and associated communities and Nifong(1998) described Carolina Bay vegetation community composition across the entiregeographic range of Carolina Bays. These studies provided important descriptiveinformation but lack quantitative data necessary for measuring or gauging the success ofrestoration activities. In addition, previous studies did not provide values that indicatethe range of vegetation communities and/or their rate of occurrence on a given soil type.Our study provides data that allows site stewards to determine restoration success basedon a range of potential communities that could be established on varying depths oforganic material.

The research reported here was conducted in three unaltered Carolina Bays thatserve as reference sites for the proposed mitigation project. The objectives of thisresearch project were to 1) describe the species composition and structure of the plantcommunities; 2) describe associations of the plant community types with soil propertiesand water table regimes; 3) determine soil and/or hydrology factors that dictatevegetation community distribution; and 4) develop objectives relating to plant communitytype, soil type, soil phosphorus content, and water table depth applicable to the long termmonitoring of Carolina Bay restoration projects in this region.

REFERENCE SITES

The study was conducted in three unaltered “reference” Carolina Bays in BladenCounty, North Carolina: Causeway Bay, Charlie Long Bay, and Tatum Millpond. TheDOT restoration site, Juniper Bay, was located approximately 30 miles southwest of thereference bays in Robeson County, North Carolina. Reference bays were chosen basedon preliminary observations of typical Carolina Bay vegetation communities, lack ofindication of recent logging or draining, soil mapping units similar to those of restorationsite, proximity to the restoration site, accessibility, and property owner cooperation. Thesoutheastern middle Coastal Plain of North Carolina is warm temperate, with a meanannual temperature of 63°F and average annual precipitation of 46” (Elizabethton, NorthCarolina, Weather Station, Lock 2).

58

The three reference bays were vegetated throughout, containing no open waterzones. They collectively exhibited a total of four vegetation communities: pond pinewoodland, non-riverine swamp forest, bay forest, and high pocosin. These vegetationcommunities were distinguished based on dominance and/or abundance of dominant treespecies and classified according to Schafale and Weakley’s (1990) communitydescriptions.

The soils in the reference bays exhibited an accumulation of organic materialfrom the periphery to the center of a bay, grading from mineral soil (0-20cm depth oforganic material) into organic soils that could be divided into three categories: histic (20-40cm organic depth), shallow organic (40-80cm organic depth), and deep organic(>80cm organic depth). The soils in all three bays were mapped as Pamlico (TerricMedisapist) in the interior, a very poorly drained series containing variable depths ofhighly decomposed organic material over a sandy sediment. In Causeway Bay andCharlie Long Bay, the Pamilco soils of the interior graded into Lynn Haven series (TypicHaplaquod) on the bay periphery, a poorly drained soil overlaying sandy sediment. InTatum Millpond, soils graded outwardly from Pamlico to Croatan (Terric Medisaprist) toTorhunta (Typic Humaquept), both very poorly drained soils overlaying a loamysediment, and eventually into Lynn Haven at the bay periphery. All four soil series weretypical of the Carolina Bays in this region (Leab, 1983; Stolt and Rabenhorst, 1987).

The reference bays were depressional wetlands that receive hydrology fromrainwater, surface water runoff, and possibly groundwater inputs. All three baysmaintained a water table depth shallower than 30cm below the soil surface for at least 2weeks during the growing season (March 16 – November 14; dates correspond to 50%probability that temperatures will drop below 28°) (Leab, 1990). Average yearly rainfallbetween 1957 and 1979 was recorded at 115cm (Leab, 1990). However, the NorthCarolina drought summary report (www.hprcc.unl.edu/Nebraska/drought-summary2002.htm/#top) indicated that rainfall was below normal in 2002. Thesereference bays were positioned at an elevation of approximately 7m above sea level,while the surrounding landscape is between 8 and 12m above sea level (USGSTopographic Map, Elizabethtown Quadrangle, 1984). In addition to rainwater andsurface water runoff, these depressions may have acted as flow-through wetlands wheregroundwater input is a major factor in their hydrology. All three bays are located in theCape Fear River watershed.

Causeway Bay was a 600 acre wetland located on private property approximately10 miles north east of Elizabethtown (34°29’N, 78°24’W). Although in the past,Carolina Bays typically experienced wildfire every 20-50 years (Christenson et al. 1988),historical aerial photographs reveal no wildfire events within the last 65 years inCauseway Bay. Similarly, these photographs indicate that the study area was not loggedwithin the same historical time span.

Charlie Long Bay, a 500 acre bay, was located on Bladen Lakes State Forestproperty (34°36’N, 78°33’W). According to retired forest service workmen, the bay was

59

never logged (Carson Tatum and J.B. Johnson, personal communication). Likewise, nowildfire events have occurred in the past 65 years, according to review of historical aerialphotographs. This bay had a surface water outlet in the northwest corner, which wasdammed in the late 1960’s for farming operations.

Tatum Millpond, approximately 1400 acres, was located 2 miles south of CharlieLong Bay on Bladen Lakes State Forest property (34°42’N, 78°33’W). This bay waslogged selectively for pond pine and Atlantic White Cedar under forest servicesupervision between 1938 and 1954. In 1954, Hurricane Hazel damaged the majority ofthe desirable/valuable trees in Tatum Millpond, and therefore, logging operations ceasedfollowing this storm event (Carson Tatum and J.B. Johnson, personal communication).

MATERIALS AND METHODS

During the summer of 2001, NCSU students/technicians established a transectfrom the periphery (wetland boundary) toward the center of each reference bay.Transects began at a point along either the northeast or southwest side of the bay(depending on accessibility), and extended into the center of the bay at an azimuth thatformed a right angle to the tangent along the bay periphery. Sampling vegetation,hydrology, and soil along these transects took place during the summer of 2002.

Sample Plots

5X5 m plots were established every 30 meters along the left side of each transectbeginning 6m from the bay periphery. The plots were distributed along each transectuntil the depth of organic material began to level. Causeway Bay and Charlie Long Baycontained 15 sample plots each, reaching 456m from the periphery toward the center ofeach bay. Tatum Millpond contained 20 sample plots, reaching 606m from the peripherytoward the center of the bay (Figures 1, 2, and 3). In March 2002, the peripheralvegetation of Tatum Millpond was clear-cut by the Forest Service. This portion (first36m; first 2 sample plots) of the transect was relocated further southeast within the bay inan area where vegetation and soils appeared similar to that portion of the originaltransect.

Vegetation Assessment

Vegetation sampling consisted of a plant inventory, species' aerial coverdetermination, shrub stem counts, and determination of tree species' density and basalarea. The species inventory, cover determination, and stem counts took place within the5X5m plots. Identification of all species found within each plot occurred via meandersearch, and a cover value was assigned to each species based on the percentage of the5X5m surface area shadowed by the projection of all parts of a given species(percentages recorded to the nearest 5%; values less than 5% were assigned a realnumber, 1 to 4%, based on estimated cover).

Shrub stems were counted within 2 two-dimensional planar bands along the northside of each 5 X 5m plot. Tallying total number of stems that passed through a horizontal

60

5m (long) X 20cm (deep) band along the soil surface of a plot resulted in anapproximated 1m2 vertical stem count for a given plot. Assessment of horizontalstructure was accomplished using the same 20cm planar band, only erecting it verticallyand tallying all vegetation parts that passed through the 3m(high) X 20cm(deep) plane at1m increments along the 5m side.

The Bitterlich method (resulting in variable plot size) was implemented throughthe use of a cruise angle for determining tree basal area from a given plot (Grosenbaugh1952). Basal area data was recorded by species from each plot, and then summed todetermine total basal area for each plot. The distance between the plots and the relativelysmall diametered trees encountered in this study ensured that trees could not be talliedtwice from two separate plots.

Soil Analysis

Soil samples were collected from a 90cm core at each 5X5m vegetation plot.Samples were collected from the 0-15cm, 25-40cm, 60-75cm, and 75-90cm depthintervals. Samples were dried, ground, and shipped to the North Carolina Department ofAgriculture, Agronomic Division, for analysis. Using a mehlich-3 extraction/mehlichbuffer acidity (Mehlich 1984), quantities of available phosphorous, potassium, calcium,magnesium, sodium, manganese, zinc, and copper were determined for each sample (4samples/plot by depth). In addition, each sample was analyzed for pH, base saturation,humic matter content, cation exchange capacity, and weight/volume. We conducted thesoil texture analysis using a standard hydrometer procedure (Ketter et al., 2001). TheNorth Carolina State University Analytical Service Laboratory completed the percentcarbon analysis utilizing continuous flow isotope ratio mass spectrometry (Goodman,1998). In addition, a complete soil profile description was recorded at each plot. Thisincluded field identification of soil texture, color, and the observed depth of organicmaterial.

Water Table Monitoring

A total of 12 wells were distributed along the sampling transects within the bays,resulting in hydrology monitoring of four soil categories: mineral soils (0-20cm depth oforganic material), soils exhibiting a histic epipedon (20-40cm depth of organic material),shallow organic soils (40-80cm depth of organic material), and deep organic soils(>80cm depth of organic material) (Figures 1, 2, and 3). The wells recorded water tabledepths on an hourly basis and allowed for the generation of seasonalwatertable/hydrographs (Fig. 4).

Data AnalysisThe plant community in each plot was classified as pond pine woodland,

nonriverine swamp forest, bay forest, or high pocosin according to the system of Schafaleand Weakley (1990) (Table 1; Figures 1, 2, and 3). Plots maintaining a dominance ofpond pine (Pinus serotina Michx.) (basal area between 40 and 60ft2/acre) were classifiedas pond pine woodland. Plots maintaining a dominance of swamp gum (Nyssa bifora

61

Walt.) (basal area between 60 and 140ft2/acre) were classified as nonriverine swampforest. Plots exhibiting a dominance of loblolly and sweet bay (Gordonia lasianthus (L.)Ellis and Magnolia virginiana L.) (combined basal area between 20 and 50ft2/acre) wereclassified as bay forest. Plots with fewer pond pine trees (less than 30ft2/acre) wereclassified as high pocosin.

Soil type (mineral, histic, shallow organic, or deep organic) of each plot wasdetermined based on percent organic matter at a given depth. Plots were classified basedon our field texture assessment, and verified using % organic carbon (OC) values. Plotslocated on soils containing <20% OC at a depth of 20-40cm were classified as mineralsoils. If soils contained >20% OC at a depth of 20-40cm, but <20% at a depth of >40cm,the plot was considered to be located on a histic epipedon. Plots with soils containing>20% OC at 40-80cm depths but <20% organic matter at depths >80cm were classifiedas shallow organic . Plots with deep organic soils maintained an organic matterpercentage of >20% at depths >80cm (Fig. 1, 2, and 3). Because the amount of clayparticles present in these soils was insignificant (<10% consistently), it was not necessaryto consider the % clay values when determining whether the plot was located on mineralor histic soils (USDA-NRCS 1998).

Subsequent to classifying the plant community and soil of each plot, the 50 plotswere grouped by soil type and plant community type. Data were analyzed with plantcommunity type as the independent variable and then again with soil type as theindependent variable. Analyses involved determining average vegetation, soil, and watertable variables associated with each plant community and soil class. The sampling designemployed here was limited by the thick, tangled, jungle-like, impenetrable vegetation thatcomprise the communities found in Carolina Bays of the North Carolina coastal plain.Therefore, although our sample size was limited and our sampling strategy representedpseudo-replication (non-random replication, therefore statistically dependent treatments),we ran several single factor analysis of variance tests (ANOVAs) in order to inferpotential significant differences between soil types or between community types,completely aware of the weakness in our sampling strategy for applying these tests(Hulbert 1984). We also tested the independence of community type to soil type utilizingcontingency table analyses and chi-squared tests to determine potential dependence ofcommunity type to soil type, aiming to reject the hypothesis that these community typesare randomly distributed in Carolina Bays.

RESULTS

Associations of Plant Community Types and Soil Types

Of the 50 plots, 13 were found on mineral soil, 13 on soils containing a histicepipedon, 6 on shallow organic soil, and 18 on deep organic soil. Of these same plots, 21were found to exhibit a pond pine woodland community, 8 exhibited nonriverine swampforest, 5 exhibited bay forest, and 16 exhibited a high pocosin community (Table 1).Table 2 indicates that mineral soil supported almost exclusively pond pine woodland,deep organics supported nearly exclusively bay forest and high pocosin, shallow organics

62

supported pond pine woodland and swamp forest, and histic soils supported a variety ofcommunity types.

We conducted a contingency table analysis for determining the interdependenceof community type on soil type. The actual values (Table 2) were compared to expectedvalues using a chi-squared test. Our The chi-squared test of independence for eachcommunity type across all soil types resulted in a values ranging between 0.08 to 0.001with 3 degrees of freedom and tabulated p-values all greater than 0.25.

Vegetation

Six tree species were identified in the reference bays: red maple (Acer rubrumL.), Atlantic white cedar (Chamaecyparis thyoides (L.) B.S.P.), loblolly bay, sweetbay,swamp gum, and pond pine. Mineral and hsitic soils exhibited relatively similar totalbasal areas (71 and 73 ft2/acre). Shallow organic soils exhibited highest total basal area(100 ft2/acre), and deep organic soils maintained the lowest total basal area (28 ft2/acre)(Table 3).

Although red maple basal area was evenly distributed (5-13 ft2/acre) between soiltypes, pond pine basal area was higher on soils with shallower organic material (54 and40 ft2/acre), swamp gum basal area was highest in shallow organic soils (70 ft2/acre), andthe basal area of bay trees combined (loblolly and sweet bay) was highest in deep organicsoils (10 ft2/acre). Atlantic white cedar was a minor component of the canopy on deeporganic soils (Table 3).

Pond pine basal area was highest in pond pine woodland (40 - 80 ft2/acre). Plotsclassified as high pocosin contained only scattered pond pine trees (less than 30 ft2/acre).Swamp gum basal area was highest in swamp forest (60 -140 ft2/acre). The combinedbasal area of loblolly bay and sweet bay was greatest in bay forest (20 - 50 ft2/acre).Red maple basal area was evenly distributed across community types (8-15 ft2/acre) withthe exception of high pocosin (2 ft2/acre). Atlantic white cedar was a minor componentof the canopy in the bay forest community (Table 4).

Table 5 combines the information obtained by grouping the plots into soil andcommunity classes. This table includes the average basal areas of dominant tree speciesfound on different soil types within different community types. For example, based onour sampling strategy and averaging analyses, pond pine woodland on mineral soil wascomprised of 58 ft2/acre pond pine, 14 ft2/acre red maple, and 3 ft2/acre bay trees. Thesevalues can be used for comparison between the same community type found on differentsoil types as well as for comparison between the same soil type that exhibits differentplant communities. By summing species’ basal area across community types within agiven soil type, we generated average basal area figures for a given community of aspecified soil type. Nonriverine swamp forest maintained the highest average total basalarea on shallow organic soils (118 ft2/acre), whereas pond pine woodland maintained thehighest average total basal area on histic soils (82 ft2/acre). High pocosin on mineral soil

63

was found to have higher average total basal area (30 ft2/acre) than pond pine woodlandfound on deep organic soils (15 ft2/acre).

A total of 31 shrub species were identified in the reference bays. The 11 mostcommon species (occurring in greater than 50% of the plots identified as a specificcommunity type, with an absolute cover value of >5%) included giant cane (Arundinariagigantea (Walt.) Muhl.), coastal sweetpepper bush (Clethra alnifolia L.), bluehuckleberry (Gaylussacia frondosa (L.) Torr. and Gray ex Torr.), large gallberry (Ilexcoriacea (Pursh) Chapman), smooth winterberry (Ilex laevigata Pursh) Gray), coastaldoghobble (Leucothoe axillaris (Lam.) D. Don), swamp doghobble (Leucothoe racemosa(L.) Gray), fetterbush lyonia (Lyonia lucida (Lam.) K. Koch), red bay (Persea palustris(Raf.) Sarg.), laurel greenbriar (Smilax laurifolia L.), and highbush blueberry (Vacciniumcorymbosum L.). Table 6 provides average aerial cover values for all dominant shrubspecies in all combinations of soil type and plant community type encountered in thisstudy. For example, in pond pine communities on mineral soil, fetterbush lyoniaaveraged 24% of the shrub coverage; however, fetterbush lyonia increased to an average37% cover when found in pond pine communities on shallow organic soil. Likewise,fetterbush lyonia cover increased to 57% of the total cover in bay forest communities,which are only found on deep organic soils.

The average stem densities of different plant community types are shown onTable 7. Pond pine woodland, nonriverine swamp forest, and bay forest exhibited similaraverage vertical stem counts per m2 (35, 36, and 33 counts, respectively). High pocosinexhibited the highest average number of vertical stem counts per m2 (52 counts). On ahorizontal plane, pond pine woodland and nonriverine swamp forest exhibited fewer stemcounts (121 and 127 counts, respectively) in a one m2 by 3m high area than bay forest andhigh pocosin (85 and 183, respectively). For comparison purposes between communities,we ran single factor analyses of variance (ANOVA) to determine significant differencesbetween the structures of the four community types. The ANOVA resulted in a p-valueof 0.015 when analyzing the vertical stem counts, indicating significant structural.However, the horizontal stem counts did not result in a p-value that indicated significancebetween community types (p-value>0.05). Unlike the structural density numbersresulting from a between community comparison, the structural density of vegetationbetween soil types resulted in similar numbers across the four soil types. Likewise, asingle factor ANOVA test using the soil classification scheme resulted in no significantdifference (p-value>0.05).

Soils

Tables 8, 9, and 10 present average values for phosphorous, potassium, calcium,magnesium, sodium, manganese, zinc, and copper for plots on different soil types withindifferent community types. There was no data available for the one plot found onmineral soil and in a high pocosin community because the test results for the samplesfrom this plot were extremely high, and therefore, thrown out. Although our sample sizewas limited and our sampling strategy represented pseudo-replication, we ran severalsingle factor ANOVAs in order to infer potential significant differences in phosphorous

64

levels between soil types and community types, completely aware of the weakness in oursampling strategy for applying these tests (Hulbert 1984). We performed single factorANOVAs to test the null hypothesis that the mean available phosphorous found betweenthe four soil and community types are equal. All tests (i.e. available phosphorousbetween soil types across bays, between soil types within individual bays, betweencommunity types across bays, between community types within individual bays) resultedin insignificant differences (p-value<0.05) except when testing variance of the meanphosphorous levels between the four soil types in Tatum Millpond (p-value=0.007).

Other than depth of organic material, no trends or significant relationships weredetected between soil type and community when examining the other soil propertiesanalyzed for this study. Cation exchange capacity fell between 8.3 and 11.1 meq/100cm3

across soil types; soil pH across soil types was between 3.5 and 3.9; and base saturationwas between 6.3% to 9.9%. Additional information regarding soil properties analyzed bydepth for the purposes of this study is available from the authors.

Hydrology

Although our hydrology monitoring did not yield testable results for the purposesof scientific analyses, it did provide information useful in making inferences that supportcurrent theories regarding the relationship between depth of organic material andhydrologic regime. Figure 3 illustrates the water table depths in Tatum Millpondbetween March 6, 2001, and August 8 2001, across all four soil categories, as an exampleof variations in hydrologic regime between the soil types found in the reference bays. Itis expected that mineral soils exhibit a deeper water table than histic soils, but shallowerthan the organic soils (Mitch and Gosselink, 1993). According to Figure 3, the watertable remained at or near the soil surface for most of March and April. With the onset ofsummer and evapotranspiration activity increasing, the water table continues to dropbelow the soil surface as the growing season begins, and remains below the soil surfacethrough out the remainder of the season, with occasional replenishment by storm events.In mineral soils, the water table dropped to nearly 60cm below the soil surface, butaveraged –33.7cm across the growing season. The watertable sampled in histic soilsdropped to 45cm below the soil surface, and averaged –19.7cm across the growingseason. The watertable in shallow organic soils fell slightly shallower (-40cm) than thewater table exhibited in histic soils, and averaged –17.9cm across the growing season.The watertable in deep organic soils dropped to 45cm below the soil surface, andaveraged 24.9cm across the growing season. Addition to water table depths, thehydrographs shown illustrated the storm events across the four soil types. They indicatethat as the depth of organic material increases, stormwater is retained for a longer periodof time, resulting in a more highly curved hydrograph rather than the sharp rises and fallsin water table depth as seen in mineral and histic soils.

DISCUSSION

Carolina Bays are considered nutrient poor systems, with deficiencies in nitrogenand phosphorous limiting productivity (Waldbridge, 1991; Richardson and

65

Vaithiyanathan, 1995; Walbridge and Richardson, 1991). Nutrient availability in soils iscorrelated with hydrologic regime. Relative to drier soils, saturated and inundated soilsare anaerobic for extended periods and this slows the decomposition of organic materialand results in a nutrient poor rhizosphere. Therefore, saturated and inundated systems areprone to organic material accumulation due to the decreased rate of decomposition.Likewise, nutrient availability decreases with an increase in organic materialaccumulation associated with longer hydroperiods and shallower water table depths(Daniel, 1982; Mitsch and Gosselink, 1993; Cronk and Fennessey, 2001). Decreasednutrient availability would result in lower primary productivity, restraining the growth ofvegetation and potentially dictating vegetation community change as available nutrientsdecrease with increased organic material accumulation (Bridgham et al., 1995; Readerand Stewart, 1972; Bridgham et al., 1991). Over the course of this study, we sought todetermine if there was a detectable association between available nutrient amount andplant community distribution in Carolina Bays in order to identify target communities forrestoration at the NCDOT – Juniper Bay wetland mitigation site. In general, we foundno significant differences in nutrient availability between soil types or between soils thatsupported different plant community types. The soils in Tatum Millpond were theexception to this general result. At this site there were significant differences in availablephosphorous across soil types and community types. This result was due to the fact thatTatum Millpond was the only reference bay that contained non- riverine swamp forest, acommunity comprised of deciduous species which return greater amounts of nutrients inleaf litter than do evergreen species. This greater nutrient return resulted in highernutrient concentrations in the organic soils of the swamp plant community than in theother soils of this site. In addition, our ANOVA’s were based on a limited sample sizeand pseudo-replication. Therefore, the results are useful for gleaning preliminary insightinto ecological trends. Likewise, it should be noted that the values obtained from thereference bays for available phosphorous and other major and minor elements allrepresented extremely low quantities when compared to more eutrophic systems.

The data obtained from the reference bays can be applied to the restoration ofdisturbed Carolina Bay systems in the North Carolina coastal plain based on soil typessimilar to those in the reference bays. The chi-squared test of independence was used todetermine significant associations between community type and soil type. The test wasapplied to contingency tables of expected number of plots verses actual number of plotsexhibiting a particular community type across soil types. We generated expected valuestables based on the ratio of the product of row totals and column totals to the grand total.We used the chi-squared test to generate a statistical value that can be tabulated toformulate a probability value that indicates significance. The chi-squared test wasapplied to each community type across soil types, and each resulted in an insignificant p-value (>0.25). An insignificant difference between the actual data and the expected dataimplies our observed values are similar to expected values. Therefore, we can infer thatthe community types identified in this study are not randomly distributed across soiltypes.

This reference data will be useful in long-term monitoring of plant communitydevelopment at restoration sites. Restoration site monitoring should note whether or not

66

the restored plant communities on a particular soil type meet the descriptions of thereference communities on comparable soil types. To facilitate the application of thereference data for monitoring purposes, we first describe the community types, theunderlying soil and water table parameters that contribute to the maintenance of thesecommunities, and other possible variables that may influence the distribution of thesecommunities in the reference bays. Secondly, we apply this information to the NCDOT –Juniper Bay wetland mitigation project in order to construct recommendations for theplanting plan and long-term monitoring of this and other potential Carolina Bayrestoration projects in the North Carolina coastal plain.

Plant Community Type Descriptions

The four plant community types identified in this study include pond pinewoodland, non-riverine swamp forest, bay forest, and high pocosin. These communitiesdiffer in canopy species composition, canopy basal area, shrub stem density, andcomposition and/or abundance of some shrub understory species. Community types aredescribed below.

Pond pine woodland was found in all three reference bays, predominantly on thewet mineral soil of the bay periphery and well represented on histic soil adjacent andtransitioning to more substantial depths of organic accumulation (Tables 1 and 2). Thewater table associated with this soil type exhibited a maximum depth of –45cm, andexperienced sharp rises and quick drainage during storm events. This hydroperiodminimizes the length of saturation and/or inundation through out the rooting zone,optimizing nutrient availability for plant uptake. In the reference bays, this communitytype exhibited an average basal area of pond pine of 58 ft2/acre on mineral soils and 68ft2/acre on histic soils (Table 5). The understory was dominated by large gallberry andfetterbush lyonia (Table 6). The shrub layer was approximately 3m tall, exhibiting adensity of an average 35 vertical stems/m2 and 121 horizontal stems/m2X3m high (Table7). The vertical density is similar to that found in the nonriverine swamp forest and bayforest communities, but significantly different than the structure exhibited by thehighpocosin community. The horizontal density is similar to that of nonriverine swamp forest,but somewhat less than bay forest and high pocosin structural architecture. Thesedistinguishing features of the pond pine woodland community type identified in thereference bays fit the description provided by Shafale and Weakley’s (1990). Themineral soils contained an average of 13.0-13.8 mg/dm3 phosphorous within a 0-90cmdepth (Table 8).

The typical pond pine woodland community is sustained by fire (Schafale andWeakley, 1990). Pond pine reproduction occurs via seed germination following fire,which is required by their serotinous cones for triggering seed release. Likewise, firereturns nutrients to the soil and maintains a relatively open understory. Because thereference bays have not been burned in at least the past 60 years, the identified pond pinewoodland communities most likely contained denser, taller shrubby vegetation than whatwould be expected of bays that have experienced wildfire once every 20-50 years, the

67

historically natural fire frequency period for Carolina Bays of this area (Richardson,1991).

Nonriverine swamp forest was found in Tatum Millpond on histic and shalloworganic soils, but was absent from the other reference bays (Tables 1 and 2). Thiscommunity type is known to occur on peatlands with seasonally saturated or inundatedsoils, as evidenced in the hydrographs (Fig. 3). .The community identified in TatumMillpond was selectively logged for Atlantic white cedar prior to 1954. Nonriverineswamp forest is known to occur in a mosaic with Atlantic white cedar forest (Schafaleand Weakley, 1990). The removal of the Atlantic white cedar trees from this systemcould have allowed nonriverine swamp forest to dominate. The reduction of Atlanticwhite cedar in the canopy and absence of fire enhanced the establishment of swamp gumand impeded the reestablishment/regeneration of the cedar (Levy and Walker, 1979;Wells and Whitford, 1976). The nonriverine swamp forest identified in Tatum Millpondcontained a closed canopy of swamp gum (87-105 ft2/acre basal area) with no co-dominant canopy species (Table 5). The understory was dominated by coastalsweetpepper bush (RIV 36%) and highbush blueberry (RIV 16%). The shrub layer alsocontained fetterbush lyonia (RIV 8%) and red bay (RIV 7%) (Table 6). The understoryexhibited an average stem density of 36 vertical stems/m2 and 127 horizontal stems/m2 x3m high (Table 7). The vertical density is similar to that found in the nonriverine swampforest and bay forest communities, but significantly different than that exhibited by thehigh pocosin community. The horizontal density is similar to that of nonriverine swampforest, but somewhat less than bay forest and high pocosin structural architecture. Thisnonriverine swamp forest adheres to the description given by Schafale and Weakley(1990). The histic and shallow organic soils contained an average of 14.9-19.9 mg/dm3

phosphorous within a 0-90cm depth (Table 8).

Historically, this community is not maintained by fire. However it is believedthat these systems receive mineral input from groundwater or surface water sources(Schafale and Weakley, 1990). Aerial photographs did not indicate a major source orhydrologic discharge point for surface water input or output at Tatum Millpond, and ourstudy does not investigate hydrology dynamics. Therefore, further research is necessaryin order to determine the hydrologic parameters necessary for maintaining thiscommunity at Tatum Millpond.

A bay forest community was identified in Tatum Millpond, but was absent fromthe other reference bays (Tables 1 and 2). This community type was found solely ondeep organic soils in the interior of the bay, where the water table exhibited long-timeretention of storm water and a watertable no deeper than 45cm below the soil surface.The canopy is dominated by bay trees (loblolly bay and sweet bay), exhibiting acombined basal area of 34 ft2/acre (Table 5). The understory was fetterbush lyonia (RIV57%), laurel greenbriar (RIV19%), and large gallberry (RIV 12%). No herbaceous layerwas present (Table 6). The understory exhibited a stem density of 33 vertical stems/m2

and 185 horizontal stems/m2X3m high (Table 7). The vertical density is similar to thatfound in the pond pine woodland and nonriverine swamp forest communities, butsignificantly different than that exhibited by the high pocosin community. However, the

68

horizontal density is high, more closely resembling the figures of a high pocosincommunity. This indicates that the structural architecture of a bay forest community ismore highly branched on a horizontal plane, but maintains a similar vertical stem densitywhen compared to pond pine woodland and nonriverine swamp forest. The observedcommunity adheres to the description given by Schafale and Weakley (1990). The histicand shallow organic soils contained an average of 12.1 mg/dm3 phosphorous within a 0-90cm depth (Table 8).

Bay forests are a component of the successional stages in peatland communitydevelopment. However, the circumstances under which a bay forest develops areagreeably unclear. A combination of fire suppression, relatively higher nutrientavailability (relative to pocosins), absence of disturbance, and presence of deep organicmaterial accumulation (deeper than a nonriverine swamp forest) are the parameters thatmay give rise to a bay forest community (Richardson, 1991; Otte, 1981; Buell and Cain,1943). In Tatum Millpond, the bay forest is found at the most centralized point in the bayalong the sampling transect. High pocosin is found between nonriverine swamp forestand bay forest; however, this is possibly due to tree falls within the identified highpocosin community associated with hurricane winds (Figure 3).

High pocosin was identified in all three reference bays, predominantly on deeporganic soils where the water table exhibited long-time retention of storm water and awater table at or above 45cm below the soil surface (Tables 1 and 2). This communitywas characterized by scattered pond pines (7-20 ft2/acre) (Table 5), with a thick/denseshrub layer growing up to approximately 3m in height and exhibiting a stem verticaldensity of 52 stems/m2 and 183 horizontal stems/m2 x 3m high (Table 7). The verticaldensity is significantly different than that of pond pine woodland, nonriverine swampforest, and bay forest communities. The horizontal density is high, similar to thatexhibited in the bay forest. The understory was dominated by fetterbush lyonia (RIV27%), highbush blueberry (RIV 15%), and large gallberry (RIV 13%) (Table 6). Thesedistinguishing features fit the description of high pocosin provided by Shafale andWeakley (1990). The deep organic soils contained an average of 8.3 mg/dm3

phosphorous within a 0-90cm depth (Table 8).

Although the reference bays have not experienced a wildfire in over 60 years, thiscommunity type is typically maintained by periodic fire (every 20-50 years). Fire returnsnutrients to the soil, encouraging high productivity and a quick reestablishment of theshrub layer. In addition, fire reduces the density and height of shrubby vegetation,maintaining a low pocosin community (Richardson, 1991). It is possible that the highpocosin communities encountered in the reference bays may indeed represent a historiclow pocosin community that has experienced fire suppression.

Planting Recommendations and Long-Term Monitoring

The NCDOT – Juniper Bay wetland mitigation site is currently a retiredagricultural field that was cleared of vegetation, ditched, and drained in the early 1970s.

69

The bay was then farmed in row crop for the subsequent 30 years. Currently, this baycontains wet mineral soil, histic soil, and shallow organic soil. For the purposes of thisstudy, we will assume the water table will be restored to the appropriate depths formaintaining the existing soil types. Due to erosion, subsidence, and a lowered watertable resulting from the initial conversion of Juniper Bay from wetland to farmland, muchof the organic material accumulation that once made Juniper Bay a peat filled bay hasdisappeared. The process of organic material accumulation will be reestablished with arestored water table. Saturation and/or inundation of the rooting zone will result inslower decomposition rates, nutrient deficiencies, and accumulation of organic material.

Ecological factors identified in association with the community types describedfrom the three reference bays are outlined in Table 11. Based on this reference data, werecommend planting the mitigation site in pond pine woodland (Fig. 4). Pond pinewoodland dominated mineral soils and was well represented on histic soils in thereference bays. Although nonriverine swamp forest dominated the shallow organic soils,this community type would not be appropriate for restoration at Juniper Bay due to 1) theunknown nature of the hydrologic regime necessary for maintaining its species structureand composition; and 2) its uncommon occurrence as a dominant community typewithout the typical Atlantic white cedar forest community mosaic dynamic. Because wehave no reference data that includes this natural mosaic pattern, we are limited to simplyexcluding nonriverine swamp forest from the recommended planting plan. Both bayforest and high pocosin would not be appropriate for restoration at Juniper Bay due totheir preference/requirement for deep organic soils that are not present at Juniper Bay. Itis likely that over time, the pond pine woodland established on shallow organic soils willaccumulate organic material due to an elevated water table, resulting in decrease innutrient availability. Consequentially, primary productivity would reduce resulting instunted shrub growth and a decreased pond pine basal area, developing into a highpocosin community (Richardson 1991). With a prescribed fire (every 20-50 years), thehigh pocosin would eventually succeed into low pocosin (Otte 1981).

Long-term vegetation monitoring should consist of a canopy and understoryspecies inventory, with the understanding of the types of communities that may evolvefrom the planted community over time. Likewise, shrub stem density and basal areashould be estimated. In addition, soil properties should be monitored to determinesuccess of nutrient level restoration to values at or near those of reference bays. Inaddition, continued hydrology monitoring in the reference bays and the mitigation bay isrecommended.

CONCLUSION

Because of the extremely dense structure of bay vegetation, we stumbledupon many sampling constraints over the course of this study. Randomly samplingthroughout the reference bays was precluded by the extreme density of the vegetation.Therefore, this research resulted in a general description of bay plant communities andthe associated soil and water table properties that probably contribute to theirdevelopment and maintenance. This baseline data will be useful to those working in the

70

field of wetland restoration in Carolina Bays of the North Carolina coastal plain. Inaddition, this research can easily be expanded upon by closer, more in depth studiespertaining specifically to hydrologic parameters that are, most likely, the ultimate dictatorof soil properties, which are, in turn, influenced by the various vegetation communitiesthat cycle nutrients through the rooting zone. Likewise, detailed studies comparingcommon disturbance regimes (flooding, fire, logging, tree fall, etc.) in this region mayalso result in significant findings pertaining to vegetation community distribution inCarolina Bays. At a regional level, these bays are valuable resources for stormwaterretention, carbon storage, and provision of unique floral habitat for wildlife. They alsorepresent a large fraction of the intact natural areas found across a highlydeveloped/agricultural landscape (Richardson et al. 1981). Therefore, further studies arenecessary for better understanding the means by which restoration of these increasinglyuncommon bay wetlands can be achieved successfully.

ACKNOWLEDGMENTS

We wish to thank the following people: Dr. Stephen Broome and Dr. JimGregory, North Carolina State University, aided us in developing a methodology thatwould result in a meaningful and applicable descriptive study; Michael Shafale, NorthCarolina Natural Heritage Program, provided data pertaining to the natural history of thereference Carolina Bays used in this study, as well as general data pertaining to CarolinaBay ecosystems of this region; Michael Chestnut, Forest Supervisor at Bladen LakesState Forest, worked with us to provide unlimited access to research sites and aided in thecollection of historical information regarding forest services practices in the past; andAlex Adams, Justin Ewing, Sara Lughinbul, Alex Baldwin, and Tripp Cox, students andtechnicians with North Carolina State University, donated countless field days for cuttingand maintaining trails, and assisted with data collection and analyses. We also wish tothank the North Carolina Department of Transportation, who fully funded this research.

71

REFERENCES

Bridgham, Scott D., and C.J. Richardson. 1993. Hydrology and nutrient gradients inNorth Carolina peatlands. Wetlands 13:207-218.

Bridgham, Scott D., C.J. Richardson, E. Maltby, and S.P. Faulkner. 1991. Cellulosedecay in natural and disturbed peatlands in North Carolina. Journal ofEnvironmental Quality 20:695-701.

Bridgham, Scott D., John Pastor, C.A. McClaugherty, and C.J. Richardson. 1995.Nutrient-use efficiency: a litter fall index, a mode, and a test along a nutrient-availability gradient in North Carolina Peatlands. American Naturalist 145:1-21.

Buell, Murray F., and R.L. Cain. 1943. The successional role of southern white cedar,Chamaecyparis thyoides, in southeastern North Carolina. Ecology 24: 85-93.

Christensen, N.L., R.B. Burchell, A.L. Liggett, and E.L. Simms. 1981. The structure anddevelopment of pocosin vegetation. p.43-61. In C. Richardson (ed.) Pocosinwetlands: An integrated analysis of coastal plain freshwater bogs in North Carolina.Hutchinson Ross, Stroudsburg, Pennsylvania.

Christensen, N.L., R.B. Wilburn, J.S. McLean. 1986. Soil-vegetation correlations in thepocosins of the Croatan National Forest, North Carolina. Unpublished report toNational Ecology Center – U.S. Fish and Wildlife Service, Fort Collins, CO, andU.S Department of the Interior – Fish and Wildlife Service, Washington, D.C.

Clark, M.K., D.S. Lee, and J.B. Funderburg. 1985. The mammal fauna of Carolina Bays,pocosins, and associated communities in North Carolina: an overview. Brimelyana11:1-38.

Cronk, J.K, and M.S. Fennessy. 2001. Wetland plants: biology and ecology. LewisPublishers. New York, NY.

Daniel, Charles C. 1981. The structure and development of pocosin vegetation. p.68-108. In C. Richardson (ed.) Pocosin wetlands: An integrated analysis of coastalplain freshwater bogs in North Carolina. Hutchinson Ross, Stroudsburg,Pennsylvania.

Grosenbaugh, L.R. 1952. Plotless timber estimates – new, fast, easy. Journal of Forestry50:32-37.

Goodman, K.J. 1998. Hardware modifications to an isotope ratio mass spectrometercontinuous flow interface yielding improved signal, resolution, and maintenance.Analytical Chemistry. 70:833:837.

Hulbert, Stuart H. 1984. Pseudoreplication and the design of ecological fieldexperiments. Ecological Monographs 54:187-211.

72

Ketter, T.A., J.W. Doran, and T.L. Gilbert. 2001. Simplified method for soil particlesize determination to accompany soil-quality analyses. Soil Science Society ofAmerica Journal. 65:849-852.

Kologiski, R.L. 1977. The phytosociology of the Green Swamp, North Carolina. NorthCarolina Agriculture Experiment Station Technical Bulletin No. 250, Raleigh, NC.

Leab, Roger J. 1990. Soil Survey of Bladen County, North Carolina. United StatesDepartment of Agriculture, Soil Conservation Service.

Levy, G.F., and S.W. Walker. 1979. Forest dynamics in the Dismal Swamp of Virginia.In P.W. Kirk (ed.) The Great Dismal Swamp. Old Dominion University ResearchFoundation, Inc.

Malmer, N. 1986. Vegetational growth in relation to environmental conditions innorthwestern European mires. Canadian Journal of Botany 64:375-383.

Mehlich, A. 1984. Mechlich 3 soil test extractant: A modification of the Mehlich 2extractant. Communications in Soil Science and Plant Analysis. 15:1409-1416.

Mitsch, W.J., and J.G. Gosselink. 1995. Wetlands. Van Nostrand Reinhold, New York,NY.

Nifong, T.D. 1998. An ecosystem analysis of Carolina Bays in the Coastal Plain of theCarolinas. Ph.D. Dissertation. University of North Carolina, Chapel Hill, NorthCarolina.

Otte, L.J. 1981. Origin, development and maintenance of pocosin wetlands of NorthCarolina. Unpublished report to the North Carolina Natural Heritage Program.North Carolina Department of Natural Resources and Community Development,Raleigh, NC.

Prouty, W.F. 1952. Carolina Bays and their origin. Bulletin of the Geological Societyof America 63:167-224.

Reader, R.J., and J. M. Stewart. 1972. The relationship between net primary productionand accumulation for a peatland in southeastern Manitoba. Ecology 53:1024-1037.

Reese, Rachel E., and K.K. Moorhead. 1996. Spatial characteristics of soil propertiesalong an elevational gradient in a Carolina Bay wetland. Soil Science Society ofAmerican Journal 60:1273-1277.

Richardson, Curtis J. 1991. Pocosins: An ecological perspective. Wetlands 11:335-354.Richardson, Curtis J., R. Evans, and D. Carr. 1981. Pocosins: An ecosystem intransition. p.3-19. In C. Richardson (ed.) Pocosin wetlands: An integrated analysisof coastal plain freshwater bogs in North Carolina. Hutchinson Ross, Stroudsburg,Pennsylvania.

73

Richardson, Curtis J., and P. Vaithiyanathan. 1995. Division S-10-Wetland Soils:Phosphorous sorption characteristics of Everglade soils along a eutrophicationgradient. Soil Science Society of America Journal 59:1782-8.

Schafale, M.P., and A.S. Weakley. 1990. Classification of the Natural communities ofNorth Carolina. North Carolina Natural Heritage Program, Division of Parks andRecreation. Department of Environment, Health, and Natural Resources, Raleigh,NC.

Schafale, M.P., and A.S. Weakley. 1991. Classification of pocosins of the CarolinaCoastal Plain. Wetlands 11:355-375.

Schalles, John F., and Donald J. Shure. 1989. Hydrology, community structure, andproductivity patterns of a dystrophic Carolina Bay wetland. Ecology 59:365-385.

Sharitz, R.R. and J.W. Gibbons. 1982. The ecology of southeastern shrub bogs(pocosins) and Carolina Bays: a community profile. U.S. Fish and WildlifeService, Division of Biological Services, Washington, D.C.

Stolt, M.H., and M.C. Rabenhorst. 1987. Carolina Bays on the eastern shore ofMaryland: I. Soil characterization and classification. Soil Science Society ofAmerica Journal 51:394-399.

USDA-NRCS. 1998. Field indicators of hydric soils in the United States, version 4.0 G.Hurt, P. Whited, and R.F. Pringle (Eds.) USDA-NRCS, Fort Worth, TX. In S.W.Sprecher. Basic concepts of soil science. In J.L. Richardson and M.J. Vepraskas.2001. Lewis Publishers, New York, NY. p 5.

Vitt, D.H., and W.L. Chee. 1990. The relationships of vegetation to surface waterchemistry and peat chemistry in fens of Alberta, Canada. Vegetation 89:87:106.

Walbridge, Mark R. 1991. Phosphorous availability in acid organic soils of the lowerNorth Carolina Coastal Plain. Ecology 72:2083-2100.

Waldbridge, Mark R., and C.J. Richardson. 1991. Water quality of pocosins andassociated wetlands of the Carolina coastal plain. Wetlands 11:417-439.

Weakley, A.S., and S.K. Scott. 1982. Natural features summary and preserve design forCarolina Bays in Bladen and Cumberland Counties, NC. Unpublished report to NCNatural Heritage Program, The Nature Conservancy, and Earthlines, Inc., Raleigh,NC.

Wells, B.W., and L.A. Whitford. 1976. History of stream-head swamp forests, pocosins,and savannahs in the southeast. Journal of the Mitchell Society 1976: 148-151.

74

Zeveloff, Samuel I. 1983. Island biogeographic considerations for pocosin wildlifeconservation. The Journal of Elisha Mitchell Scientific Society 99:69-77.

75

Table 1. Number of plots found on each soil type and within each community type bybay.

Reference Bays

Soil/Community TypeCauseway

BayCharlie

Long BayTatum

MillpondTotal no.of plots

Mineral (<20cm OM) 4 7 2 13Histic (20-40cm OM) 2 7 4 13Shallow organic (40-80cm OM) 1 1 4 6Deep organic (>80cm OM) 8 0 10 18

Pond Pine Woodlanda 5 13 3 21Nonriverine Swamp Forest 0 0 8 8Bay Forest 0 0 5 5High Pocosin 10 2 4 16

Total # of plots 15 15 20 50aSchafele and Weakley (1990).

76

Table 2. Percentage (absolute values in parentheses) of total plots found in pond pinewoodland, nonriverine swamp forest, bay forest, and high pocosin communities on thespecified soil types.

Community typePond PineWoodland

NonriverineSwamp Forest Bay Forest

HighPocosin

Soil Type -------------------------------------%---------------------------------Mineral (N=13)a 92 (12)b 0 0 8 (1)

Histic (N=13) 54 (7) 23 (3) 0 23 (3)

Shallow Organic (N=6) 33 (2) 67 (4) 0 0

Deep Organic (N=18) 0 6 (1) 27 (5) 67 (12)a Number of plots sampled.b Values indicate the specified communities are found independently on a particular soil type; not to beinterpreted that community types are found intermingled on a particular soil type.

77

Table 3. Average absolute basal area (ft2/acre) and average relative basal area (inparentheses, ft2/acre) of dominant tree species within the specified soil types.

Soil Type

SpeciesMineral(N=13)a

Histic(N=13)

ShallowOrganic

(N=6)

DeepOrganic(N=18)

Red Maple 13 ± 9b

(18)min=0

max=30

12 ± 10(16)

min=0max=30

12 ± 8(12)

min=0max=20

5 ± 7(18)

min=0max=30

Atlantic white cedar 0 1 ± 3(1)

min=0max=10

0 2 ± 7(6)

min=0max=20

Loblolly bay 3 ± 5(4)

min=0max=10

1 ± 1(1)

min=0max=5

0 7 ± 15(26)

min=0max=50

Sweet bay 0 0 2 ± 4(2)

min=0max=10

3 ± 6(12)

min=0max=20

Swamp gum 0 20 ± 38(27)

min=0max=90

70 ± 57(70)

min=0max=140

7 ± 15(23)

min=0max=60

Pond Pine 54 ± 17(77)

min=20max=70

40 ± 29(54)

min=0max=80

17 ± 26(17)

min=0max=50

4 ± 9(16)

min=0max=30

Average Total 71 73 100 28a Number of plots sampled.b Standard deviation.

78

Table 4. Average absolute basal area (ft2/acre) and average relative basal area (inparentheses, ft2/acre) of dominant tree species within the specified community types.

Community Type

Species

Pond PineWoodland(N=21)a

NonriverineSwampForest(N=8)

Bay Forest(N=5)

High Pocosin(N=16)

Red maple 15 ± 8b

(20)min=0

max=30

8 ± 7(7)

min=0max=20

12 ± 8(23)

min=0max=20

2 ± 3(13)

min=0max=10

Atlantic white cedar 5 ± 2(1)

min=0max=10

0 6 ± 13(12)

min=0max=30

0

Loblolly bay 3 ± 3(2)

min=0max=10

0 26 ± 19(50)

min=0max=50

1 ± 3(6)

min=0max=10

Sweet bay 0 1 ± 4(1)

min=0max=10

8 ± 8(15)

min=0max=20

1 ± 3(8)

min=0max=10

Swamp gum 0 93 ± 24(87)

min=60max=140

0 4 ± 7(21)

min=0max=20

Pond Pine 59 ± 13(77)

min=40max=80

5 ± 10(5)

min=0max=30

0 9 ± 10(53)

min=0max=30

Total Average 76 106 52 17a Number of plots sampled.b Standard deviation

79

Table 5. Average basal area (ft2/acre) of Red maple (RM), Swamp gum (SG), Pond pine (PP), and the bay trees (BT) across mineral,histic, shallow organic, and deep organic soil types within pond pine woodland, nonriverine swamp forest, bay forest, and highpocosin community types.

Community Types

Pond Pine WoodlandNonriverine

Swamp Forest Bay Forest High PocosinSoil Type RM SG PP BT RM SG PP BT RM SG PP BT RM SG PP BTMineral (N=13)a 14 0 58 3 -- -- -- -- -- -- -- -- 0 0 20 10

Histic (N=13) 19 0 63 0 3 87 13 0 -- -- -- -- 0 0 20 2

Shallow Organic(N=6)

15 0 50 0 10 105 0 3 -- -- -- -- -- -- -- --

Deep Organic(N=18)

-- -- -- -- 10 60 0 0 12 0 0 34 2 5 7 2

a Number of plots sampled

80

Table 6. Average absolute vegetation cover (% cover) of all common shrub species (RIVin parentheses) across mineral, histic, shallow organic, and deep organic soil types withinpond pine woodland, nonriverine swamp forest, bay forest, and high pocosin communitytypes.

Community Types

Soil TypePond PineWoodland

NonriverineSwamp Forest Bay Forest High Pocosin

Mineral Giant cane Coastal sweetpepper Blue huckleberry Large gallberry Smooth winterberry Coastal doghobble Swamp doghobble Fetterbush lyonia Red bay Laurel greenbriar

Highbush blueberry

6 (7)7 (3)8 (7)

30 (32)3 (3)

00

17 (24)1 (1)3 (3)

10 (11)

----------------------

----------------------

1 (1)1 (1)

10 (11)30 (32)

000

40 (44)00

10 (11)

Histic Giant cane Coastal sweetpepper Blue huckleberry Large gallberry Smooth winterberry Coastal doghobble Swamp doghobble Fetterbush lyonia Red bay Laurel greenbriar

Highbush blueberry

13 (21)6 (8)1 (1)

27 (27)4 (4)

00

16 (25)1 (1)2 (3)6 (8)

0 (0)10 (3)1 (1)

15 (13)2 (2)

01 (1)

31 (53)5 (7)3 (3)

14 (16)

----------------------

7 (6)7 (6)

17 (17)25 (15)5 (5)

00

18 (17)1 (1)3 (3)

17 (15)

Shallow Organic Giant cane Coastal sweetpepper Blue huckleberry Large gallberry Smooth winterberry Coastal doghobble Swamp doghobble Fetterbush lyonia Red bay Laurel greenbriar

Highbush blueberry

2 (2)5 (7)4 (4)

22 (24)8 (9)

00

28 (37)2 (2)1 (2)

15 (16)

030 (36)

00

2 (2)5 (6)

10 (11)9 (8)8 (7)5 (6)

14 (16)

----------------------

----------------------

81

Deep Organic Giant cane Coastal sweetpepper Blue huckleberry Large gallberry Smooth winterberry Coastal doghobble Swamp doghobble Fetterbush lyonia Red bay Laurel greenbriar

Highbush blueberry

----------------------

010 (11)

000

10 (11)10 (11)5 (6)4 (4)

025 (53)

05 (4)

014 (12)

000

69 (57)5 (4)

23 (19)0

6 (5)9 (8)

12 (7)18 (13)1 (1)

00

35 (27)3 (3)

15 (3)22 (15)

82

Table 7. Average structural density by community type based on counts derived fromnumber of times vegetation (leaf, twig, branch…) passes through a 3-D vertical andhorizontal 20cm band along a 5m transect.

Community TypeVertical Stem Counts Horizontal Stem

Counts

Pond Pine Woodland 35 121

Nonriverine Swamp Forest 36 127

Bay Forest 33 185

High Pocosin 52 183

83

Table 8. Average measured quantities of phosphorous (mg/dm3) at a 0-90cm depthacross Mineral, Histic, Shallow Organic, and Deep Organic soil types within Pond PineWoodland, Nonriverine Swamp Forest, Bay Forest, and High Pocosin communities.

Community Types

Soil TypePond PineWoodland

NonriverineSwamp Forest Bay Forest High Pocosin

Mineral 13.0 -- -- --

Histic 13.8 14.9 -- 4.3

Shallow Organic 10.0 19.9 -- --

Deep Organic -- 15.9 12.1 8.3

84

Table 9. Average measured quantities (meq/dm3) of potassium (K), calcium (Ca), magnesium (Mg), and sodium (Na) at a 0-90cmdepth across Mineral, Histic, Shallow Organic, and Deep Organic soil types within Pond Pine Woodland, Nonriverine Swamp Forest,Bay Forest, and High Pocosin communities.

Community Types

Pond Pine WoodlandNonriverine

Swamp Forest Bay Forest High PocosinSoil Type K Ca Mg Na K Ca Mg Na K Ca Mg Na K Ca Mg NaMineral 0.07 0.29 0.19 0.05 -- -- -- -- -- -- -- -- -- -- -- --

Histic 0.09 0.20 0.23 0.09 0.12 0.23 0.22 0.17 -- -- -- -- 0.15 0.23 0.51 0.09

Shallow Organic 0.11 0.27 0.21 0.09 0.12 0.22 0.24 0.36 -- -- -- -- -- -- -- --

Deep Organic -- -- -- -- 0.10 0.18 0.23 0.25 0.15 0.22 0.77 0.32 0.12 0.21 0.54 0.15

85

Table 10. Average measured quantities (meq/dm3) of manganese (Mn), zinc (Zn), and copper (Cu) at a 0-90cm depth across mineral,histic, shallow organic, and deep organic soil types within pond pine woodland, nonriverine swamp forest, bay forest, and highpocosin community types.

Community Types

Pond Pine WoodlandNonriverine

Swamp Forest Bay Forest High PocosinSoil Type Mn Zn Cu Mn Zn Cu Mn Zn Cu Mn Zn CuMineral 1.21 0.78 0.11 -- -- -- -- -- -- -- -- --

Histic 1.15 1.74 0.06 1.17 2.16 0.14 -- -- -- 0.79 1.89 0.05

Shallow Organic 0.98 1.48 0.06 1.14 2.76 0.10 -- -- -- -- -- --

Deep Organic -- -- -- 0.93 3.00 0.03 1.11 4.13 0.11 1.06 5.79 0.05

86

Table 11. Suggested ecological parameters for monitoring purposes in the establishmentand/or restoration of pond pine woodland, nonriverine swamp forest, bay forest, and highpocosin communities in Carolina Bay ecosystems, based on the ecological parametersexhibited by the three reference Carolina Bays sampled over the course of this study.

Ecological Parameters Pond PineWoodland

NonriverineSwamp Forest

Bay Forest High Pocosin

DominantCanopy Speciesand Basal Area

Pond Pine58-68 ft2/acre

Swamp Gum87-105ft2/acre

Bay Trees34

ft2/acre

Pond Pine7-20

ft2/acre

DominantUnderstoryComposition

large gallberry

fetterbushlyonia,

coastalsweetpepper bush

highbushblueberry

fetterbushlyonia

laurelgreenbriar

largegallberry

fetterbushlyonia

highbushblueberry

largegallberry

Vertical Density(Understory Stems)

Horizontal Density(Understory Stems)

35/m2

121/m2

X 3m high

36/m2

127/m2

X 3m high

33/m2

185/m2

X 3m high

52/m2

183/m2

X 3m high

Soil Type Mineral Histic HisticShallow Organic

Deep Organic Deep Organic

Available Phosphorous 13.0-13.8mg/dm3

14.9-19.9mg/dm3

12.1mg/dm3

8.3mg/dm3

Average WaterTable Depth

-33.7cm -19.7cm -17.9cm -24.9cm

87

0m 50 100 150 200 250 300 350 400 450 500

cm20406080

100 Depth of Organic Material120

Pond PineWoodland

HighPocosin

Min

eral

His

tic

Dee

pO

rgan

ic

Sha

llow

Org

anic

Figure 1. Plot layout along transect at Causeway Bay from periphery(0m) to center (500m) of the bay. = plot; • = hydrology monitoring well.

88

Figure 2.Plotlayoutalongtransect atCharlieLongBayfromperiphery(0m)tocente

r (500m) of the bay. = plot; • = hydrology monitoring well.

0m 50 100 150 200 250 300 350 400 450 500

cm2040 Depth of Organic Material 6080

100120

Pond Pine Woodland High

PocosinPond Pine Woodland

Min

eral

His

tic

Min

eral

His

tic

Sha

llow

Org

anic

89

0m 50 100 150 200 250 300 350 400 450 500

cm 20 40 60 80 100 120 Depth of Organic Material

140

PondPineWoodland

NonriverineSwamp Forest

High Pocosin Bay Forest

Min

eral

His

tic

Dee

pO

rgan

ic

Sha

llow

Org

anic

Figure 3. Plot layout along transect at Tatum Millpond from periphery (0m) toward thecenter of the bay. Circled area indicates portion of transect relocated due to clear-cuttingof original transect in those areas. = plot; • = hydrology monitoring well.

90

Deep Organic Hydrograph

March April May June July August

30cm

0

-30

-60

March April May June July August-60

-30

0

3030cm

0

-30

-60

Shallow Organic Hydrograph

March April May June July August

Histic Hydrograph

March April May June July August

Mineral Hydrograph

0

30cm

30cm

-30

-30

0

-60

-60

Figure 4. Hydrographs for March 6, 2001 through August 9, 2001 from mineral, histic,shallow organic, and deep organic soils in Tatum Millpond.

91

= Shallow Organic

= Mineral

= Histic

Pond Pine Woodland

Figure 5. Recommended planting plan for specified soil types at the Juniper Bay DOTmitigation/restoration site

92

Chapter 5

CHEMICAL SOIL PROPERITES OF JUNIPER BAY AFTER 15, 20, AND 30YEARS OF DRAINAGE AND AGRICULTURAL PRODUCTION

J.M. Ewing and M.J. Vepraskas

INTRODUCTION

Section 404 of the Clean Water Act (1977) requires the replacement, ormitigation, of destroyed wetlands. Wetland mitigation includes, enhancement andpreservation of current wetlands, creation of new wetlands, or the restoration of priorwetlands (USCOE, 2002). Success of such mitigation projects has been mixed. Erwin(1991) found that of 40 mitigation projects in south Florida 60% were judged incompleteor failures, Wilson and Mitsch (1996) found that only 38% of the desired wetland wasestablished at mitigation sites in Ohio, and Gallihugh and Rogner, (1998) found that 99ha of 128 mitigation sites involving 144 ha, were found to have unsatisfactory hydrology.The probability of success should be greater in an area that once supported wetlandhydrology, hydric soils, and hydrophytic vegetation.

Drained Carolina bays found along the Atlantic Coast in the southeastern U.S.,are being utilized for wetland restoration. Carolina bays are elliptical depressions in thelandscape that are orientated along the long axis SE to NW. They range in size from 10m to >4 km along the long axis. The bays are usually surrounded by a light coloredsandy rim and have a dark colored depression resulting from high amounts of organicmatter (Johnson, 1942). The extent of these bays range from Northern Florida toDelaware with the highest concentration in North and South Carolina. Estimates on thenumber of these bays are as high as 500,000 (Johnson, 1942), but the actual numbermaybe less than 100,000 (Nifong, 1998). Many theories for bay formation have beenproposed, from the most popular, meteor impact (Johnson, 1936), artesian springs(Prouty, 1952; LeGrand, 1952), whale wallows (Grant 1945), and ice flows (Bliley andBurney, 1988). Currently the most plausible explanation is that originally there wereshallow depressions in the landscape with an aquitard that allowed the water table to beheld above the surface. Prevailing winds then shaped the depression into the nowfamiliar orientated shape (Thom, 1970; Odum, 1952). During the past centuryagricultural and community development have led to the drainage and use of these bays.It is estimated that 50% of all Carolina bays were drained and developed in some mannerin Bladen County, NC by 1982 (Weakley and Scott, 1982). This figure would be higherif other management practices such as logging were included. As these bays are used foragriculture and other activities, their defining characteristics of sand rims and organicsurfaces, become blurred into the surrounding landscape.

Plant communities typically found in undrained Carolina bays include non-riverine swamp forest, low pocosin, high pocosin, pond pine woodland, peatland AtlanticWhite Cedar forest, and Bay forest (Schafale and Weakly, 1990). These plant

93

communities are found in nutrient poor soils that maybe organic soil or mineral soils.Variability in these plant communities depends on depth of organic matter, seasonalwater table depths, fire and mineral input if any (Schafale and Weakly, 1990). Soil seriesassociated with theses plant communities (Croatan, Pamlico, Ponzer, Lynn Haven,Torhunta, Rutlege, and Pantego) are very strongly acidic (4.5-5.0) to ultra acidic (<3.5).Cation exchange capacity tends to be very low ranging from 1 to 30 cmol kg-1, but can beas high as 100 cmol kg-1 in the surface layers. Exchangeable phosphorus has found to bestrongly limiting to plant production in High Pocosin plant communities (Schafale andWeakly, 1990).

The North Carolina Department of Transportation (NCDOT) purchased a 256 hadrained Carolina bay in 1999, near Lumberton, North Carolina. This land was purchasedwith the intent of earning wetland mitigation credits. The Carolina bay being restored bythe NCDOT is called Juniper Bay. Juniper Bay was drained, cleared, and put intoagricultural production incrementally in 1971, 1981 and 1986. A review of the history ofthe clearing process was described by Ewing et al. (2003). Agricultural practices in theorganic soils include the addition of lime to achieve a pH of 5.5 to 5.0 (Lilly and Baird,1993). To reach this target pH up to 13470 kg ha-1 may have to be applied to overcomethe large reserve of acidic cations present in organic soils (Lilly, 1981). The NorthCarolina Cooperative Extension Service recommends the application of 135-180 kg N ha-

1, 35-55 kg P2O5 ha-1, and 90-115 kg K ha-1 for corn production (Crozer, 2000). Organicsoils in North Carolina have been shown to leach nitrogen and phosphorus and have to beapplied yearly (Lilly, 1981). Crop cultivation, depending on the crop, can include the useof chisel plows, moleboard plows, disks, and subsoilers. The overall objective of theJuniper bay project is to restore Juniper bay back to an ecosystem that is typical ofCarolina bays. This includes restoration of hydrology, vegetation, and soils typicallyfound in Carolina bays.

The objective of this study was to evaluate the changes in soil chemical propertiescreated by agricultural practices in Juniper Bay. Changes resulted from the addition offertilizer and lime over 30 years as well as tillage and drainage of the area.

MATERIALS AND METHODS

Juniper Bay is located 10 km southeast of Lumberton, North Carolina in RobesonCounty (34o30’30”N 79o01’30”E). The Bay was logged, drained, and put intoagricultural production in three stages. The western third of Juniper Bay was drained byditches in 1971, the central third and most of the eastern third was drained in 1981, thearea to the north in the eastern third was drained in 1986. A perimeter ditch was dug in1971 around the entire bay, two main drainage ditches that run NE to SW and manylateral ditches that run NW to SE. The perimeter ditch and the main ditches wereapproximately 7 m wide and 4 m deep. The lateral ditches were roughly 1.5 m wide and1m deep. Areas that are enclosed by ditches are called ‘cuts’. There was only one outletin Juniper Bay and it is located at the SW end of the main ditch on the western side. Asoil survey of Juniper Bay indicated three broad groups of soil that are based on the

94

thickness of the organic layer. They were mineral (<20cm organic material), histicepipedon (20-40 cm organic material), and organic (>40 cm organic material).

Sampling locations were chosen by randomly placing an equilateral triangle gridover a map of Juniper Bay. Grid points were spaced far enough apart to allow 24sampling points to be within the site boarded by the perimeter ditch, with ten located inorganic soil, eight in mineral soil, and six in soils with a histic epipedon. Those pointswere found in the field using GPS. Using the location point from the grid as a reference,we dug a 1 to 1.5 m pit at the center of a field cut and a pit near the closest lateral ditch,resulting in a pair of pits at each location. Soil profiles were described and bulk samplesfor laboratory analysis were taken from each horizon.

Three reference bays were selected in Bladen County, North Carolina based ontheir similarity of soils to Juniper Bay, lack of drainage or agricultural applications, andownership cooperation. Three reference sites were named Charlie Long-Millpond Bay(34o46’00”N 78o33’30”E), and Tatum Millpond Bay (34o43’00”N 78o33’00”E), bothlocated in the Bladen Lakes State Park and Causeway Bay (34o39’45”N 78o25’45”E),which was located 10 km N of White Lake, N.C. A trail from the rim of each bay to thecenter was cut through the dense vegetation. The three broad soil groups described inJuniper Bay, mineral, histic, and organic, were also found in all three reference bays.Plots were marked off at 50m intervals along the transect. Plots to be sampled were thenrandomly selected in each soil. There were a total of eight sampling plots in eachreference bay, two in the mineral and histic soils, and four in the organic soils. Soilprofiles in the selected plot were described from samples extracted with a McCauley peatsampler and an open bucket auger. Bulk samples were taken for laboratory analysis fromeach horizon.

Bulk samples from Juniper Bay and the reference bays were air dried and groundwith an electric grinder to pass through a 2-mm sieve. Extractable K, Ca, Na, Mg, Mn,Zn, and Cu were determined by running Mehlich III extract (Mehlich, 1984) through aninductively coupled plasma emission spectrograph. Cation exchange capacity and sum ofbase cations were also determined (Mehlich, 1976). The pH was determined using a 1:1soil to water ratio. Organic carbon and total nitrogen were determined through drycombustion with a Perkin-Elmer PE2400 CHN Elemental Analyzer (Culmo, 1988).

Soil types, organic, histic, and mineral, were analyzed separately using the SASprocedure PROC MIXED (SAS, 1985), with an AR(1) covariance structure to compareyears since drainage and proximity to a ditch. The number of locations for a given soiltype and years in agricultural production can be seen in Table 6. Each location hadpaired pits with one located at the crest of the cut and one near the ditch. The horizontype and depth present at each location varied and were evaluated as a spatially repeatedmeasure. The data were not balanced so least square means (LSMEANS) were used toobtain estimated means. Comparisons among crest and near ditch pits were made withp>0.1 being significant.

95

The comparisons between the reference bays and Juniper Bay were analyzedseparately depending on, soil type, organic, histic, and mineral, using the SAS procedurePROC MIXED (SAS, 2000), with an AR(1) covariance structure. Data from the crestlocations in Juniper Bay were used in this analysis after determining that differences increst and ditch locations were minimal. Organic, histic, and mineral soils were combinedinto one data set among the three reference bays, resulting in 12 organic, six histic, andsix mineral locations. Horizons varied depending on location and pit and were evaluatedas a spatially repeated measure. The reference Carolina bays were considered randomvariables so results could be inferred to natural Carolina bays as a whole. Soils from thereference bays were assumed to be equal to the Juniper Bay soils prior to drainage andchemical application. Data were not balanced so LSMEANS were used to obtainestimated means. Comparisons among Juniper Bay and the reference bays and time sincedrainage were made with p>0.1 being significant.

RESULTS AND DISCUSSION

Typical profile descriptions for organic, histic, and mineral soils from Juniper Bayand the reference bays are given in Table 1. The soils in Juniper Bay have been affectedby tillage and drainage. The surface horizons do not have an organic mat, Oi or Oehorizons that are typically found in undrained Carolina bays. The surface horizons in JBhave moderate to strong structure while the RB have weakly developed structure.Drainage has allowed the deeper organic soil in JB to develop structure while similar soilin the RB is massive. The agriculture producer who owned the land prior to the NCDOThad soil fertilizer records from 1976, which showed one to three tons/ac of dolomiticlime, 50 to 200 lb P2O5, and 200 to 300 lb K/ ac was applied. Mr. Freeman also indicatedthat fertilizer recommendations were followed yearly.

ORGANIC SOILS

Juniper Bay vs. Reference Bays

There were significant differences in organic carbon, extractable K, Ca, Mn, BS,CEC, and total N between the organic soils of Juniper Bay and the organic soils of thereference bays (Table 2). The organic soils in the reference bays had 8.4 g kg-1 more OCat the surface than in Juniper Bay (Fig. 2a). The RB soils decreased in OC content withdepth, 37.6 g kg-1 at the surface to 3.7 g kg-1 at 175 cm. The change from organic soil tomineral soil in the RB occurred around 150 cm. The OC increased in JB from 29.2 g kg-1

at the surface to 35.2 g kg-1 at 32 cm, and then decreased with depth to 2.5 g kg-1 at 108cm. This shows that oxidization at the surface was reducing the soil OC content, butbelow the plow layer the rate of oxidation was slower and OC was similar to that of theRB. The change from organic soil to mineral soil in JB occurred at approximately 75 cm.Total nitrogen followed a similar trend as OC and was higher through the profile in RB.Total N content in general, corresponds with the amount of organic matter in the soil,however it does not relate to the amount available to plants (Tonapa, 1974). Nitrogenlevels in JB were 0.87 g kg-1 at the surface, decreasing to 0.5 g kg-1 at 108 cm, and in RBthey decreased from 1.46 g kg-1 at the surface to 0.07 g kg-1 at 173 cm.

96

Organic soils have very high CEC’s, which in general increases as OM increases(Kamprath, and Welch, 1962). Such is the case in JB where the CEC is lower throughoutthe profile ranging from 24.0 cmolc kg-1 at the surface to 4.0 cmolc kg-1 at 108 cm whilein the RB it ranged from 46.6 cmolc kg-1 at the surface to 6.6 cmolc kg-1 at 171cm. Thismeans higher amounts of lime are required to change pH values because of the largereserves of acidic cations (Juno and Kamprath, 1979) and the ability to adsorb largeamounts of basic cations from lime (Evans and Kamprath, 1970). Such was the case inJuniper Bay where agricultural producers added several tons of lime per ha each toincrease and maintain an elevated pH (Freeman, 2000 personal interview). The pH in JBwas 0.72 to 0.24 units higher through the profile than in RB. The pH in both decreasedwith depth through the organic material and then increased when mineral material wasencountered. Due to the high amounts of lime applied extractable Ca was higher in thesurface soils of JB and decreased with depth to a level found in RB (Fig 2b). ExtractableCa was 4.5 cmol Ca kg-1 higher at the surface in JB and continued to be elevated overlevels in the RB to a depth of 88cm. There was a similar trend with extractable Mn was6.83 mg kg-1 higher at the surface in JB and continued to be elevated over RB levels to adepth of 75 cm. Histosols contain no or very little K-supplying or K-fixing silicateminerals and essentially all the K present is held on the OM exchange complex and isreadily avaible (Lilly, 1981). It would be expected that JB have equal or higher amountsof K due to fertilizer additions, however, extractable K was higher in the upper 20cm ofthe RB, 1.10 cmol K kg-1 compared to JB 0.33 cmol K kg-1. This might be because cropshad remove the readily avaible K. As a result of the decreased CEC and increased levelof base cations, base saturation in JB was consistently higher than the RB (Fig 2c). TheRB had a BS of 12.0% at the surface and decreased with depth while at JB it was 61.4%at the surface decreasing to 15.4% at 84cm, then increased slightly to 20.0% at 108cm.

It was expected that levels of extractable P would show signs of leaching throughthe profile in Juniper Bay because organic soils are low in Al and Fe oxides will not holdinorganic P, and can be leached from pure organic soils or quartz sand. Fox andKamprath, 1971; Larson et al., 1959). However extractable P, as well as Mg and Zn,were not significantly different but had some notable trends. Extractable P and Mgfollowed a familiar trend of having higher levels in JB and decreasing with depth tolevels found in the RB, but the opposite was true for extractable Zn, which had highervalues at the surface in the RB. The ability of organic soils to hold P will increase withtime as subsidence, ditch spoil, and lime increase the mineral content in the root zone(Larson et al., 1958), and may be the case in Juniper Bay.

Time in Agricultural Production

Nutrient levels, pH and base saturation have been reported to increase in Carolinabay soils when under agricultural production (Ewing, 2002; Hanchey et al., 2000).Increased periods of time in agricultural production resulted in significant increases of BSand extractable Ca. (Fig. 3) in the organic soils of Juniper Bay. Base saturation increasesreflect combined impacts of Ca, Mg, Na and K. Depths of apparent leaching of Ca alsoincreased with time in agricultural production. Calcium levels were increased over these

97

in the RB down to depths of approximately 40, 70, and 90 cm after 15, 20, and 30 yearsin production. This results in an estimated leaching rate of 2.5, 3.4, and 2.9 cm yr-1 for15, 20, and 30 years of agricultural production.

Crest vs. Ditch

There was a significant difference in the amount of zinc between the crest andditch locations (Fig. 1). Extractable Zn was higher at the crest than the ditch, 10.40 and5.6 mg kg-1 in the surface 15 cm respectively and decreased with depth to 0.38 and 0.22mg kg-1 at depths of 107 and 110cm. Levels of Zn reached a low constant level below1.00 mg kg-1 at 48cm at the ditch and 74cm at the crest.

Although there were no differences found in extractable P, K, Ca, Mg, Mn, CEC,BS, and pH for the organic soil there was a similar depth trend of higher levels in thesurface 15cm at the crest relative to the ditch and tended to be elevated through theprofile at the crest compared to the ditches. For example, extractable P was 81.0 mg kg-1,at the crest and 59.0 mg kg-1 at the ditch. Organic carbon was above 12 g kg-1 from 0-74cm at the crest and 0-48cm at the ditch. The trends seem to indicate that there is moreoxidation of the organic soil near the ditch due to better drainage. The trends in nutrientsindicate a continued overlap during application of fertilizer and lime, or that nutrientshave been remove near the ditch through drainage or surface runoff.

HISTIC SOILS

Juniper Bay vs. Reference Bays

LSMEANS estimates comparing Juniper Bay and the reference bays for histicsoil types can be seen in Table 3. The general trend for both JB and RB is to have thehighest amount of constituent at the surface and decrease with depth. There were twodifferent trends, one, like extractable P (Fig. 5a), had higher levels the upper 30cm at JBand then decreased to levels found in the RB, and the other, like extractable K (Fig. 5b)had higher levels in the upper 30 cm at the RB.

Base saturation, extractable Ca, Mg, Zn, followed the trend illustrated byextractable P. Extractable P levels at the surface for JB were 96.4 mg kg-1 and decreasedwith depth to levels found in RB. There was 36.6 mg kg-1 of ex. P in the surface 20cm atthe RB and decreased to 9.5 mg kg-1 at 32cm and remained constant through the rest ofthe profile. Extractable Ca in the surface 20cm at JB was 3.88 cmol Ca kg-1 and 0.43cmol Ca kg-1 at RB. Extractable Mn at the surface in JB was 8.05 mg kg-1 and of 3.37mg kg-1 at the RB. Extractable Mg at JB was 1.54 cmol Mg kg-1 in the surface 20cm atJB and 0.86 cmol Mg kg-1, at the RB. Base saturation in the surface 20cm is 63.0% at JBand 9.6% in the RB.

Organic carbon, total N, and CEC followed the trends illustrated by extractable K.Extractable K was higher in the surface 20cm at the RB 0.71 cmol K kg-1, compared toJB 0.28 cmol K kg-1. Levels below 20 cm were similar for both locations and decreased

98

with depth. The histic soils in JB had lower amounts of OC (17.0%) at the surface 20cmcompared to the histic soils at the RB (30.5%). This is reversed at 30 cm with 24.7% OCat JB and 13.5% at RB. Below 40 cm the OC levels decrease with depth for JB and RB.The organic mineral boundary in JB was at 30cm and 32cm in RB. Total N was 1.08% atthe surface in the RB and 0.51 in JB. CEC in JB is 33.05 cmolc kg-1 in the surface 20cmand decreased with depth, RB is highest at the surface, 17.83 cmolc kg-1.

Although not significant the pH in JB was highest in the surface horizon 4.29,and ranged to a low of 3.81, while at the RB the high pH, 4.37 was in the C horizon andranged to 3.51 near the surface.

Time in Agricultural Production

Differences based on the amount of time since drainage were significant forpercent organic carbon, CEC, base saturation, and extractable Ca, and Mn. The amountof organic carbon after 0, 15 and 20 years of drainage, in the surface 18cm was 30.5,20.2, and 13.4% respectively and decreased with depth (Fig. 6a). However, areas drainedfor 30 years first increased to 39.2% at 32cm, then decreased. Extractable Ca washighest, 5.08 cmol Ca kg-1 in the surface 20 cm in the areas under agricultural productionfor 20 years, followed by 2.69 cmol Ca kg-1 in the areas under production for 15 years.Time zero had 0.43 cmol Ca kg-1 or less extractable Ca throughout the profile. The15year areas decreased with depth and reached low levels (<0.20 cmol Ca kg-1) found inthe RB at 52cm, and the 20 year areas reached RB levels at 104cm. This means that Camove through the soil 2.0 and 3.6 cm yr-1 for 15 and 20 years of agricultural production.Extractable Mn followed the same trend as extractable Ca and had 10.38 mg Mn kg-1 inthe surface 20 cm in the areas under agricultural production for 20 years, and 5.71 mg kg-

1 in the areas under production for 15 years with 3.37 mg kg-1 at time zero. ExtractableMg move through the soil 2.0 and 3.6 cm yr-1 for 15 and 20 years of agriculturalproduction. CEC at time zero was higher than 15 or 20 years in the upper 50cm due tohigher amounts of organic matter, but lower below 50cm because of lower amounts clay(Fig 6b). CEC in the RB was 33.05 cmolc kg-1 in the surface 16cm, areas in productionfor 20 years had 19.36 cmolc kg-1 in the surface 19cm. Base saturation was highestthrough the profile after 20 years of production, followed by areas in production for 15years (Fig 6c). Base saturation at time zero were never above 9.6%, but ranged from52.4 to 12.9 for areas drained for 15 years and from 73.4 to 12.5% for areas drained 20years.

Crest vs. Ditch

There were significant differences between crest and ditch for extractable P, Mg,Mn, and Zn, and base saturation. Extractable P levels at the crest were 48 to 25 mg kg-1

higher at the crest compared to the ditch in the upper 30 cm (Fig 4). Extractable P levelsfor both crest and ditch dropped to low (<3.5 mg kg-1) levels below 40 cm. The sametrend of higher levels in the surface 30cm at the crest decreasing to similar low levels isevident in extractable Mg, Mn, and Zn and BS. The surface soils in the histic soil typehad 6.38 and 4.07 mg kg-1 extractable Zn at the crest and ditch. Extractable Mn was

99

higher in the surface horizon at the crest 8.07 mg kg-1 than the ditch 4.07 mg kg-1.Extractable Mg was higher at the crest 1.58 cmol Mg kg-1 in the surface horizoncompared to the ditch 0.77 cmol Mg kg-1. The base saturation was higher in the upper30cm crest surface soil by 23-11% when compared to the ditch.

Although not significant the trends of CEC, extractable K and Ca looked like thatof extractable P and tended to be higher in the surface 30 cm at the crest position anddecrease to similar low levels. The pH was similar a both locations with highest at thesurface and ranged from 3.80 to 4.88 at the crest and 4.07 to 4.68 at the ditch. Organiccarbon was above 12% in the surface 30cm at both positions. Organic carbon increasedfrom 15cm to 30 cm before decreasing below 12%. The lower levels of OC at the surfacecould be due to increased oxidation or addition of ditch spoil.

MINERAL SOILS

Juniper Bay vs. Reference Bays

LSMEANS estimates comparing Juniper Bay and the reference bays for mineralsoil types can be seen in Table 5. Levels of base saturation, pH, extractable Ca, P, andZn had higher levels in the surface 30cm in JB then decreased with depth to levels foundin the RB. This trend is illustrated with extractable P (Fig. 8a). Levels of extractable Pare 0.33 and 87.1 mg kg-1 in the surface 10cm at the RB and JB respectively.Extractable P decreases rapidly to 16.0 mg kg-1 at 44cm in JB, and remained near levelsfound in the RB for the remaining of the profile. Base saturation was highest in theupper 10cm, 16.3%, at the RB and 74.6% in JB, and then decreased with depth.Extractable Ca in the surface 10cm was 1.15 cmol Ca kg-1 in the RB and 3.34 cmol Cakg-1 in JB. Zinc levels were 12.4 and 5.96 mg kg-1 in the surface 10 cm in JB and the RBrespectively. The pH in JB in the surface 30cm was approximately 5.0 and decreased toaround 4.3 below 62cm while in the RB the pH was approximately 3.65 in the upper 25cm increasing to 4.1 below 52cm. Organic carbon, total N, CEC, extractable K and Mghad the opposite trend with higher levels in the surface layers in the RB decreasing tolevels that are similar to those in JB. This trend is illustrated with extractable K in Fig8b. Extractable K levels in the surface horizon were 0.26 and 0.75 cmol K kg-1 at JB andRB respectively. JB decreased to 0.04 cmol K kg-1 at 24cm and the RB decreased to0.14. Below this depth extractable K remained around 0.04 cmol K kg-1 for the rest ofthe profile in JB, but continued to decrease to 0.01 cmol K kg-1 in the RB. Organiccarbon in the surface 10cm of the RB is 24.9% and 7.64% in JB. The high OC in the RBis due to an accumulation of leaf litter at the surface. Total nitrogen in the surface 10cmwas higher in the RB, 0.73% than in JB, 0.19%. CEC levels in the RB were highest inthe surface horizon, 32.62 cmolc kg-1, and decreased with depth to 8.67 cmolc kg-1 at22cm. Juniper Bay soils had a CEC of 12.71 cmolc kg-1 in the surface horizon, decreasedwith depth to 2.63 cmolc kg-1 at 126cm. Extractable Mg in the surface horizon for JB was1.12 cmol Mg kg-1 and 1.23 cmol Mg kg-1 for the RB. There were no significantdifferences between JB and the RB for extractable Mn, however it did follow the trend ofhaving higher levels at the surface in JB and decreasing to levels found in the RB.

100

Time of Drainage

Significant effects due to the time of drainage and agricultural production werefound in base saturation (Fig. 9). Time zero, the reference bays, had a BS of 16.3% at thesurface and decreased with depth to 4.98 at 82cm. After 25 years of agriculturalproduction the BS in the surface 10 cm was 76.48% and decreased with depth to 27.08%at 120cm. After 30 years of agricultural production BS in the surface 10 cm was 72.76%decreased rapidly to 46.79% 23cm, and continued to decrease with depth to 25.08% at57cm before increasing to 37.43% at 132cm.

Although not significant extractable Ca, Mg, P, Mn, and Zn followed the sametrend with areas drained for 30 years having higher levels in the surface soils, closelyfollowed by areas drained for 20 years, followed by time zero levels. Levels of theseconstituents, from either drainage time, decreased with depth to levels found in the RB.In areas drained for 20 years the pH levels were higher than areas drained for 30 years,and both drainage areas were generally higher than pH from the RB. The surface horizonat time zero was above 12% organic carbon and was 10 cm thick. The area drained for30 years had 10.5 and 13.0% organic carbon in the surface two horizons or 23cm. Thesevalues may indicate that this location may use to have been a histic or organic horizon.The organic carbon in the areas drained for 20 years is less than 5% through out theprofile.

Crest vs. Ditch

Although not significant extractable P and CEC followed the trend of havinghigher levels in the surface horizons at the crest and decreasing with depth to levelssimilar to that of the ditch (Table 4). Organic carbon and total nitrogen were very similarat both locations with the highest level being at the surface and decreasing with depth.

Levels of extractable Ca, K, Mg, Mn, and Zn, pH and BS in the surface horizonwere significantly higher at the crest compared to the ditch. All of these variablesdecrease with increasing depth down to 30-50 cm where the values reach a relativelyconstant equilibrium (Fig. 7). Extractable Ca was twice as high at the crest compared tothe ditch in the surface horizon, then decreased to similar levels by 30cm and continuedto decrease through the profile. Extractable K was 0.206 cmol K kg-1 at the crest and0.071 cmol K kg-1 at the ditch in the surface horizon. Extractable Mg was 0.66 cmol kg-1

higher in the surface horizon at the crest but decreased to similar levels found in the ditchprofile by 50cm. Extractable Mn and Zn were 3.65 mg kg-1 and 6.77 mg kg-1 higher atthe crest in the surface horizon. The pH in the surface horizon at the crest was 5.04 thendecreased to 4.47 at 50cm then varied between 4.47 and 4.23 for the remained of theprofile. At the ditch pH was 4.68 in the surface horizon, then increased to 4.76 at 50cm,and then decreases to 4.23 at 125cm. Because of the elevation in Ca and Mn, basesaturation is 22.4% higher at the crest compared to the ditch.

101

CONCLUSIONS

Three reference bays were used to estimate the original conditions of the soils atJuniper Bay prior to agricultural practices being imposes. In general, soils of thereference bays had higher nutrient levels at the surface due to biocycling, however, thoselevels were significantly less than those found at the surface in Juniper Bay. The organicsoils at Juniper Bay have increased levels of extractable K, Ca, Mn, and base saturation,and decreased levels of organic carbon and total nitrogen in the surface 75 cm ascompared to the reference bays. The increases in nutrient levels and base saturation are aresult of fertilizer and lime applications from agricultural practices. The decreases inorganic carbon, total nitrogen and CEC are due to oxidation of the organic soil. Thehistic soils in Juniper Bay are higher in extractable Ca, Mg, P, Mn, and base saturation inthe surface 30 to 50cm compared to the RB as a result of agricultural additions.Extractable K, organic carbon, total nitrogen and CEC are lower in Juniper bay due tooxidization of organic matter and crop up take. The mineral soils at Juniper bay hadhigher levels of Extractable Ca, Mg, P, Zn, base saturation, and pH compared to thereference bays because of agricultural additions. There were lower levels of extractableK, organic carbon, total nitrogen, and CEC in Juniper bay because of the loss of a litterlayer and plant up take.

The differences in the soils at Juniper bay when evaluated at crest and ditchlocations showed that Extractable Zn was higher at the crest in organic soils. ExtractableMg, P, Mn, Zn and base saturation were higher at the crest in the histic soils. ExtractableK, Ca, Mg, Mn, and Zn were higher at the crest in the mineral soils. The increased levelsat the center of the field may be due to over a continual practice of overlappingapplication passes due to the spacing of the cuts. Additionally, nutrients at the ditch mayhave been lost through surface runoff or diluted by additions of spoil from ditchmaintenance. While differences exist between crest and ditch locations in each soil, theyboth are well above levels found in the reference bays and could be managed in a similarfashion.

When looking at changes over time, the trend for all the soils is for higher nutrientlevels, base saturation and pH and lower organic carbon, CEC, and total nitrogen thelonger the soil has been drained and in agricultural production. However, onlyextractable Ca and base saturation were shown to be affected in this manner in theorganic soils. The histic soils had more extractable Ca, Mn, base saturation, and lowerCEC and organic carbon after 20 years of production compared 15 years of production,which had higher levels than time zero. In the mineral soils, base saturation and CECwere the only variable that shows a correlation to the length of time in production withbase saturation increasing over time and CEC decreasing. The length of time influencedthe depth to which there were differences, i.e. the longer in production the deeper thedifferences occur.

The changes in soil chemical properties due to agricultural production are large.We believe that the increases in nutrient levels and pH will negatively influence theestablishment of plant communities typical of Carolina bays, which thrive under acidic

102

and nutrient poor conditions. Invasive plant species will thrive and out compete speciesthat are planted for restoration purposes for nutrients and light resources. To betterunderstand how these differences will affect the establishment of desired species, and tofind soil management practices that facilitate the re-vegetation efforts, further studiesshould be conducted.

103

REFERENCES

Bliley, D.J. and D.A. Burney. 1988. Late Pleistocene climatic factors in the genesis of aCarolina Bay. Southeastern Geo. 29:83-101.

Clean Water Act Section 404. 1977. [33 USC 1344]

Crozer, C. 2000. Fertilizer and lime management. North Carolina Corn ProductionGuide. North Carolina Cooperative Extension Service, North Carolina StateUniversity: Raleigh, N.C.

Culmo, R.F. 1988. Principle of Operation – The Perkin-Elmer PE 2400 CHN ElementalAnalyzer. Perkin Elmer Corp. Norwalk, CT.

Ewrin, K.L. 1991. An Evaluation of Wetland Mitigation in the Southeast. In J.A. Kuslerand M.E. Kentula eds. Wetland Creation and Restoration. Island Press,Washington DC, pp. 233-265.

Ewing, J.M., M.J. Vepraskas, and C.W. Zanner. Soil and Sediment Characteristics of aDrained Carolina Bay. ASA proceedings, 2002.

Evans, C.E. and E.J. Kamprath. 1970. Lime responses as related to percent Alsaturation, solution Al, and organic matter content. Soil Sci. Soc. Am. Proc.34:893-896.

Fox, R.L., and E.J. Kamprath. 1971. Adsorption and leaching of P in acid organic soilsand high organic matter sand. Soil Sci. Soc. Am. Proc. 18:76-79.

Freeman Jr, R. 2001, Personal Interview.

Gallihugh, J.L. and J.D. Rogner. 1998. Wetland Mitigation and 404 Permit ComplianceStudy, Vol. I. U.S. Fish and Wildlife Service, Region III, Burlington, IL, and U.S.Environmental Protection Agency, Region V, Chicago. 161pp.

Grant C. 1945. A biological explanation of the Carolina Bays. The Scientific Monthly.61:443-450.

Hanchey, M., C. J. Richardson, and N. Flanagan. 2000. A comparison of Carolina Baysoil properties under agriculture, wetland restoration and reference conditions.WetlandWire. 3:4.

Johnson, D.W. 1942. Origin of the Carolina Bays. Columbia University Press.

Johnson. D.W. 1936. Origin of the Supposed Meteorite Scars of the Carolina. Science. 84:15-18.

104

Juno, A.S.R., and E.J. Kamprath. 1979. Copper chloride as an extractant for estimatingthe potentially reactive aluminum pool in acid soils. Soil Sci. Soc. Am. J. 43:35-38.

Kamprath, E.J. and C. D. Welch. 1962. Retention and cation exchange properties oforganic matter in coastal plain soils. Soil Sci. Soc. Am. Proc. 26:263-265.

Knight, R.L., B.H. Winchester and J.C.Higman. 1985. Carolina Bays-Feasibility foreffluent advanced treatment and disposal. Wetlands. 4:177-203.

Larsen, J.E., R. Langston, and G.F. Warren. 1958. Studies on the leaching of appliedlabeled phosphorus in organic soils. Soil Sci. Soc. Am. Proc. 22:558-560.

Larsen, J.E., G.F. Warren, and R. Langston. 1959. Effect of iron, aluminum and humicacid on phosphorus fixation by organic soils. Soil Sci. Soc. Am. Proc. 23:438-440.

LeGrand, H. E. 1953. Streamlining of the Carolina Bays. J. Geol. 61:263-274.

Lilly, J.P., 1981. The Blackland Soils of North Carolina: Their Characteristics andManagement for Agriculture. North Carolina Ag. Res. Tech Bul. No. 270. pp. 27-33.

Lilly, J.P. and Baird. 1993. North Carolina Cooperative Extension Service Pub. AG439-17. North Carolina State University: Raleigh, N.C.

Mehlich A. 1984. Mehlich-3 soil test extractant: a modification of Mehlich-2 extractant.Commun. Soil Sci. Plant Anal. 15(12):1409–16.

Mehlich A. 1976. New buffer method for rapid estimation of exchangeable acidity andlime requirement. Commun. Soil Sci. Plant Anal. 7(7): 637–52.

Nifong, T.D. 1998. An ecosystem analysis of Carolina Bays in the Coastal Plain of theCarolinas. Ph.D. Dissertation. University of North Carolina, Chapel Hill, NorthCarolina.

Odum, H. T. 1952. The Carolina Bays and a Pleistocene weather map. Am. J. Sci.250:263-270.

Prouty, W.F. 1952. Carolina Bays and their Origin. Bulletin of the Geological Societyof America. 63:167-224.

Reese, R. E. and K. K. Moorhead. 1996. Spatial characteristics of soil properties alongan elevational gradient in a Carolina Bay wetland. Soil Sci. Soc. Am. J. 60:1273-1277.

105

SAS Institute Inc. 2000. SAS version 8.2. Cary, NC, USA.

Schafale, M.P. and A.S. Weakley. 1990. Classification of the Natural Communities ofNorth Carolina: Third Approximation. North Carolina Natural Heritage Program,Division of Parks and Recreation, Department of Environment, Health andNatural Resources. Raleigh, NC. p. 205-217.

Sharitz, R.R. and J.W. Gibbons. 1982. The ecology of southeastern shrub bogs(pocosins) and Carolina Bays: A community profile. Fish and Wildlife Service.

Thom, B.G. 1970. Carolina Bays in Horry and Marion Counties, South Carolina. Geo.Soc. Am. Bull. 81:783-814.

Tomapa, Sampe. 1974. Nitrogen studies in Histosols. Ph.D Thesis, North Carolina StateUniv., Raleigh, N.C.

US Army Corps of Engineers. 2002. Regulatory Guidance Letter No. 02-2.

Weakley, A.S. and S.K. Scott. 1982. Natural features summary and preserve design forCarolina Bays in Bladen and Cumberland Counties, NC. Unpublished report toNC Natural Heritage Program, The Nature Conservancy, and Earthlines, Inc.,Raleigh, NC.

Wilson, K.A. and W.J. Mitsch. 1996. Functional assessment of five wetlands constructedto mitigate wetland loss in Ohio, USA. Wetlands 16:436-451.

106

Table 1. Example profile descriptions of organic, histic, and mineral soils from thereference bays and Juniper Bay.

Horizon Depth Description

cm Reference bay Mineral

Oi 0-10 Brown (7.5YR 3/4) fibric organic material, root mat and leaf litter.

Weak medium platy structure. Clear boundary.

A 10-20 Black (10YR 2/1) sandy loam. Weak fine granular structure. Clear

boundary.

E 20-40 Gray (10YR 6/1) sand. Single grain structure. Clear boundary.

Bh1 40-70 Black (10YR 2/1) sandy loam. Weak medium sub-angular blocky

structure. Clear boundary.

Bh2 70-100 Very dark grayish brown (10YR 3/2) loamy sand. Weak coarse

sub-angular blocky structure.

JB17C Core17: Mineral

Ap 0-20 Black (10YR 2/1) loamy sand with 95% organically coated sand

grains. Weak medium (2-5mm) granular structure. Abrupt

boundary.

E 20-47 20-47cm. Gray (N 6/0) loamy sand with 12% A horizon material in

root channels. Single grain structure. Clear boundary.

Bhir1 47-70 Very dark brown (10YR 2/2) sandy loam with 25% E material in

upper half of horizon. Massive structure. Clear boundary.

Bhir2 70-92 Black (10YR 2/1) and dark brown (10YR 3/3) sandy loam in

alternating layers 2-7cm thick. Massive structure. Abrupt boundary.

107

Bh 92-120 Black (10YR 2/1) sandy loam with 10% gray (10YR 6/1) sandy

grains. Massive structure. Firm and slightly brittle.

Reference bay Histic

Oi 0-15 Very dark brown (10YR 2/1) hemic muck. Weak medium platy

structure. Clear boundary. Root mat.

Oa 15-36 Black (N 2.5/0) sapric muck. Massive structure. Clear boundary.

OC 36-65 Black (N 2.5/0) sandy loam. Massive structure. Gradual boundary.

C 65-100 65-100cm. Very dark brown (10YR 2/2) sand. Single grain

structure.

JB61C: Histic

Ap 0-22 Black (N 2.5/0) mucky sandy loam with 60% organically coatedsand grains. Moderate medium granular structure. Clear boundary.

Oa 22-36 Very dark brown (10YR 2/2) sapric muck. Strong coarse granular

structure. Clear boundary.

Bw 36-56 Dark brown (10YR 3/3) sandy loam with 2% gray (10YR 6/1) sand.

Weak medium sub-angular blocky structure. Clear boundary.

BC 56-74 Dark yellowish brown (10YR 4/6) loamy sand with 2% gray

(10YR 6/1) sand. Weak medium sub-angular blocky structure.

Gradual boundary.

C 74-108 Brownish yellow (10YR 6/6) with 2% gray (10YR 6/1) sand. Single

grain structure.

Reference bay Organic

108

Oi 0-20 Black (10YR 2/1) fibric to hemic muck that is part of the root mat.

Weak coarse platy structure from layering of plant debris. Gradual

boundary.

Oe 20-50 Very dark brown (10YR 2/2) hemic muck. Massive structure with

many organic bodies 0.5 to 1.0 cm in diameter. Gradual boundary.

Oa 50-85 Black (10YR 2/1) sapric muck. Massive structure. Gradual

boundary.

OC 85-110 Very dark brown (10YR 2/2) mucky loam or mucky sandy loam.

Massive structure. Gradual boundary.

2C1 110-140 Very dark brown (10YR 2/2) sandy loam. Massive structure.

Gradual boundary.

2C2 140-180 Dark gray (10YR 4/1) sand. Single grain structure.

JB10C: Organic

Ap 0-10 Very dark brown (10YR 2/2) sandy loam. Moderate medium (3mm)

granular structure. Abrupt boundary.

Oa 10-20 Black (N 2.5/0) sapric muck. Moderate fine (2mm) granular to sub-

angular blocky structure. Abrupt boundary.

OA1 20-51 Black (10YR 2/1) mucky silt loam. Weak very coarse (20cm)

prismatic structure. Clear boundary.

OA2 51-61 Very dark brown (10YR 2/2) mucky silt loam with 25% black (N

2.5/0) charcoal. Weak coarse (10cm) prismatic structure.

OA3 61-68 Dark brown (10YR 3/3) mucky fine sandy loam. Very weak coarse

(10cm) prismatic structure. Clear boundary.

109

Bw 68-107 Yellowish brown (10YR 5/4) very fine sandy clay loam. Weak

very coarse (20cm) platy to moderate medium (5cm) sub-angular

blocky structure. Clear boundary.

C 107-120 Light brownish yellow (10YR 6/2) sandy clay loam. Strong

medium (1cm) angular blocky structure. Faint reaction to alpha-

alpha.

110

Table 2. LSMEANS estimates for Juniper Bay (JB) and the reference bays (RB) in an organic soil. (*) Significant at p<0.1

Horizon Depth K* Ca* Mg P Mn* Zn BS* CEC* pH C N*

(cm) ---------- (cmol kg-1) --------- ---------- (mg kg-1) ----------- -- (%) -- (cmolc kg-1) ------- (%) -------

Juniper Bay

Op 15 0.329 5.16 2.01 76.9 11.44 10.41 61.37 23.97 4.43 29.23 0.83Oa1 32 0.248 2.6 1.12 32.36 5.06 3.93 34.69 21.45 4.03 35.17 0.75Oa2 54 0.167 1.41 0.58 12.31 2.91 1.56 21.89 15.63 3.75 21.49 0.5Oa3 75 0.134 0.82 0.34 7.91 2.25 0.88 15.84 13.26 3.77 16.4 0.37Bw 87 0.076 0.35 0.14 6.03 0.63 0.28 16.81 6.23 3.88 6.52 0.142C 108 0.069 0.23 0.1 4.06 0.53 0.18 20.05 4.01 3.97 2.54 0.05

Error 0.18 0.28 0.37 6.2 0.68 4.9 2.03 6.03 0.07 4.9 0.21

Reference Bays

Oi 19 1.1 0.62 1.75 70.77 4.61 15.05 11.97 46.6 3.73 37.58 1.46Oe 48 0.35 0.31 0.5 25.14 2.06 16.48 9.26 23.3 3.51 28.11 0.97

Oa1 88 0.067 0.21 0.28 10.96 1.14 7.46 5.46 17.92 3.5 20.45 0.54Oa2 123 0.009 0.09 0.07 3.87 0.4 2.56 3.27 10.7 3.87 16.1 0.34OC 148 0.021 0.14 0.1 2.67 0.65 1.35 3.52 10.48 3.95 13.66 0.272C 171 0.051 0.07 0.1 0.1 0.36 0.47 4.41 6.59 3.98 3.77 0.07

Error 0.11 0.23 0.23 5.1 0.58 3.6 2.39 3.88 0.06 3.7 0.14

111

Table 3. LSMEANS estimates for Juniper Bay (JB) and the reference bays (RB) in a soil with a histic epipedon. (*) Significant atp<0.1

horizon depth K* Ca* Mg* P* Mn* Zn BS* CEC* pH C* N*

(cm) ---------- (cmol kg-1) ---------- --------- (mg kg-1) ---------- -- (%) -- (cmolc kg-1) ------- (%) -------

Juniper Bay

Ap 18 0.28 3.88 1.54 96.4 8.05 6.34 62.96 17.83 4.9 17 0.51Oa 30 0.23 1.83 0.83 35.1 3.57 2.16 30.76 16.06 4.02 24.7 0.46

Bw1 51 0.13 0.59 0.37 1.4 1.81 0.46 16.61 10.29 3.83 10.2 0.23Bw2 65 0.08 0.21 0.1 2.2 0.46 0.28 13.74 9.88 3.81 1 0.06BC 95 0.07 0.17 0.11 0.2 0.42 0.24 20.18 3.63 3.97 0.8 0.04C 112 0.06 0.16 0.09 2.7 0.53 0.28 15.1 4.61 3.87 1 0.06

Error 0.08 0.28 0.15 5.8 0.7 1.02 2.75 2.8 0.1 3.1 0.091

Reference bays

Oi 16 0.71 0.429 0.865 36.6 3.37 6.34 9.57 33.05 3.63 30.5 1.08Oa 32 0.18 0.165 0.153 9.5 1.23 3.84 7.04 11.02 3.51 13.5 0.47OC 56 0.08 0.177 0.086 9.6 0.98 1.19 6.05 8.48 3.72 6.5 0.212C1 82 0.06 0.104 0.013 3.9 0.26 0.39 6.17 5.2 3.98 3.6 0.12C2 118 0.04 0.058 0.024 3.8 0.14 0.87 4.84 3.65 4.37 1.9 0.05Error 0.08 0.31 0.13 7.4 0.7 1.39 2.61 2.8 0.1 3.3 0.095

Table 4. LSMEANS estimates for mineral soils in a crest vs. ditch comparison. (*) Significant at p<0.1

112

Horizon depth K* Ca* Mg* P Mn* Zn* BS* CEC pH* C N

(cm) ------ (cmol kg-1) ----- ------ (mg kg-1) ----- -- (%) -- (cmolc kg-1) ----- (%) -----

Crest1 15 0.206 3.36 1.07 85.67 8.37 12.02 71.5 12.54 5.04 7.69 0.272 31 0.045 1.52 0.52 32.11 2.12 3.99 52.9 7.26 4.93 6.42 0.173 48 0.035 0.45 0.24 14.61 0.42 0.56 31.9 4.06 4.47 1.87 0.044 67 0.052 0.27 0.18 5.67 0.62 0.22 24.2 4.18 4.29 1.56 0.035 92 0.052 0.27 0.16 6.53 0.59 0.37 27.7 3.41 4.30 1.47 0.036 107 0.031 0.18 0.10 8.99 0.27 0.17 26.6 2.33 4.47 1.01 0.017 122 0.330 0.19 0.08 4.35 0.30 0.05 25.0 2.08 4.23 0.63 0.01

Error 0.028 0.31 0.11 6.17 0.61 0.99 5.50 1.20 0.17 1.80 0.04

Ditch1 15 0.071 1.76 0.39 66.51 4.48 5.25 49.1 8.57 4.68 7.54 0.252 32 0.040 1.33 0.33 27.44 2.13 3.54 45.6 7.01 4.75 6.89 0.193 49 0.037 0.65 0.18 9.47 0.92 1.09 37.2 4.61 4.76 3.91 0.064 70 0.025 0.38 0.18 14.61 0.48 0.29 29.5 4.08 4.5 1.67 0.035 89 0.033 0.24 0.13 15.17 0.34 0.21 24.5 2.97 4.38 1.21 0.026 107 0.062 0.19 0.09 10.86 0.42 0.31 16.1 3.51 4.11 1.45 0.047 125 0.050 0.14 0.07 15.14 0.31 0.45 14.5 3.00 4.04 0.91 0.02

Error 0.028 0.31 0.11 6.17 0.61 0.99 5.5 1.20 0.17 1.70 0.04

Table 5. LSMEANS estimates for Juniper Bay (JB) and the reference bays (RB) in a mineral soil. (*) significant at p<0.1

113

Horizon Depth K* Ca* Mg* P* Mn Zn* BS* CEC* pH* C* N*

(cm) ------- (cmol kg-1) ------- ------- (mg kg-1) ------ --- (%) --- (cmolc kg-1) ----- (%) -----

Juniper Bay

1 10 0.257 3.343 1.12 87.1 9.05 12.44 74.62 12.71 5.1 7.64 0.272 27 0.035 1.68 0.66 33.5 2.37 5.23 56.13 7.85 4.91 7.31 0.223 44 0.031 0.54 0.31 16 0.09 0.49 35.01 4.7 4.55 2.48 0.064 62 0.059 0.34 0.18 7.4 0.54 0.17 27.36 4.23 4.31 1.83 0.045 85 0.047 0.34 0.13 8.1 0.64 0.5 30.91 2.96 4.34 1.43 0.036 101 0.034 0.28 0.12 10.3 0.27 0.38 31.21 2.63 4.55 0.73 0.017 126 0.035 0.26 0.09 5 0.34 0.38 32.25 2.41 4.29 0.19 0

Error 0.07 0.35 0.3 5.6 2.8 1.31 4.52 3.02 0.25 3.1 0.07

Reference bays

1 10 0.75 1.15 1.23 33.3 8.69 5.96 16.3 32.62 3.74 24.95 0.732 22 0.141 0.11 0.16 8.7 1.08 0.97 8.47 8.67 3.62 6.09 0.193 52 0.001 0.03 0.04 3.1 0.26 0.49 7.8 1.71 4.07 0.26 0.014 82 0.001 0.069 0.06 8 0.27 0.24 4.98 3.58 4.12 0.96 0.02

Error 0.08 0.38 0.23 6.8 1.9 1.37 4.97 3.51 0.21 2.5 0.07

114

Table 6. Number of locations in relation to general soil type and years of agriculturalproduction.

15 years 20 years 30 years

Organic Soil 2 2 4

Histic Soil 3 3

Mineral Soil 5 4

115

0

20

40

60

80

100

120

0 3 6 9 12

Extractable Zn(mg kg-1)

Dep

th (

cm)

C

D

Figure 1. Comparison of extractable Zn at the crest (C) and ditch (D) in organic soils atJuniper Bay.

116

0

20

40

60

80

100

120

140

160

180

0 5 10 15 20 25 30 35 40

Organic Carbon(%)

Dep

th (

cm)

JB

RB

a.

0

20

40

60

80

100

120

140

160

180

0 2 4 6

Extractable Ca(cmol kg-1)

Dep

th (

cm)

JB

RB

b.

117

0

20

40

60

80

100

120

140

160

180

0 10 20 30 40 50 60 70

Base Saturation(%)

Dep

th (

cm)

JB

RB

c.

Figure 2. Comparison of organic carbon (a), extractable Ca (b), and base saturation (c),or organic soils from Juniper Bay (JB) and the reference bays (RB).

118

0

20

40

60

80

100

120

140

160

180

0 10 20 30 40 50 60 70 80

Base Saturation(%)

Dep

th (

cm)

0

15

2030

a

0

20

40

60

80

100

120

140

160

180

0 2 4 6 8

Extractable Ca(cmol kg-1)

Dep

th (

cm)

0

15

20

30

b.

Figure 3. Relation of years in agricultural production on base saturation (a) andextractable Ca (b) in organic soils.

Years

Years

119

0

20

40

60

80

100

120

0 25 50 75 100

Extractable P(mg kg-1)

Dep

th (

cm)

C

D

Figure 4. Crest (C) vs. Ditch (D) comparison of extractable P in soils with a histicepipedon at Juniper Bay.

120

0

20

40

60

80

100

120

0 20 40 60 80 100 120

Extractabel P(mg kg -1)

Dep

th (

cm)

JB

RB

a.

0

20

40

60

80

100

120

0 0.25 0.5 0.75

Extractable K(cmole kg-1)

Dep

th (

cm)

JB

RB

Figure 5. Comparisons of extractable P (a) and K (b) in soils with a histic epipedon atJuniper Bay (JB) and the reference bays (RB).

121

0

20

40

60

80

100

120

140

0 2 4 6

Extractable Ca(cmol kg-1)

Dep

th (

cm)

0

15

20

a.

0

20

40

60

80

100

120

140

0 10 20 30 40 50

CEC(cmolc kg-1)

Dep

th (

cm)

0

15

20

b.

122

0

20

40

60

80

100

120

140

0 20 40 60 80

Base Saturation(%)

Dep

th (

cm)

0

15

20

c.

Figure 6. Relationship of time in agriculture on extractable Ca (a), CEC (b), and basesaturation (c) in soils with a histic epipedon.

123

0

20

40

60

80

100

120

140

0 1 2 3 4

Extractable Ca(cmole kg-1)

Dep

th (

cm)

C

D

Figure 7. Crest (C) vs. Ditch (D) comparison of extractable Ca in mineral soils ofJuniper Bay.

124

0

20

40

60

80

100

120

140

0 50 100

Extractable P (mg kg-1)

Dep

th (

cm)

D

ND

a.

0

20

40

60

80

100

120

140

0 0.5 1

Extractable K(cmol kg-1)

Dep

th (

cm)

D

ND

b.

Figure 8. Juniper Bay (JB) verses Reference Bays (RB) on extractable P (a) and K (b) ina mineral soil.

125

0

20

40

60

80

100

120

140

0 50 100

Base Saturation(%)

Dep

th (

cm)

0

20

30

Figure 9. Time of agricultural practices on base saturation in a mineral soil

126

Chapter 6

PHYSICAL AND MORPHOLOGICAL CHARACTERISTICS OF CAROLINABAY WETLAND SOILS AFTER 15, 20, AND 30 YEARS OF DRAINAGE AND

AGRICULTURAL PRODUCTION

J.M. Ewing and M.J. Vepraskas

INTRODUCTION:

Section 404 of the Clean Water Act (1977) requires the replacement, ormitigation, of destroyed wetlands. Wetland mitigation includes, enhancement andpreservation of current wetlands, creation of new wetlands, or the restoration of priorwetlands (USCOE, 2002). Success of such mitigation projects is mixed. Erwin (1991)found that of 40 mitigation projects in south Florida 60% were judged incomplete orfailures, Wilson and Mitsch (1996) found that only 38% of the desired wetland wasestablished at mitigation sites in Ohio, and Gallihugh and Rogner, (1998) found that 99ha of 128 mitigation sites involving 144 ha, were found to have unsatisfactory hydrology.The probability of success should be greater in an area that once supported wetlandhydrology, hydric soils, and hydrophytic vegetation.

Drained Carolina bays found along the Atlantic Coast in the southeastern U.S.,are being utilized for wetland restoration. Carolina bays are elliptical depressions in thelandscape that are orientated along the long axis SE to NW. They range in size from 10m to >4 km along the long axis. The bays are usually surrounded by a light coloredsandy rim and have a dark colored depression resulting from high amounts of organicmatter (Johnson, 1942). The extent of these bays range from Northern Florida toDelaware with the highest concentration in North and South Carolina. Estimates on thenumber of these bays are as high as 500,000 (Johnson, 1942), but the actual numbermaybe less than 100,000 (Nifong, 1998). Many theories for bay formation have beenproposed, from the most popular, meteor impact (Johnson, 1936), artesian springs(Prouty, 1952; LeGrand, 1952), whale wallows (Grant 1945), and ice flows (Bliley andBurney, 1988). Currently the most plausible explanation is that originally there wereshallow depressions in the landscape with an aquitard that allowed the water table to beheld above the surface. Prevailing winds then shaped the depression into the nowfamiliar orientated shape (Thom, 1970; Odum, 1952). During the past centuryagricultural and community development have led to the drainage and use of these bays.It is estimated that 50% of all Carolina bays were drained and developed in some mannerin Bladen County, NC by 1982 (Weakley and Scott, 1982). This figure would be higherif other management practices such as logging were included. As these bays are used foragriculture and other activities, their defining characteristics of sand rims and organicsurfaces, become blurred into the surrounding landscape.

Plant communities typically found in undrained Carolina bays include non-riverine swamp forest, low pocosin, high pocosin, pond pine woodland, peatland AtlanticWhite Cedar forest, and Bay forest (Schafale and Weakly, 1990). These plant

127

communities are found in nutrient poor soils that maybe organic soil or mineral soils.Variability in these plant communities depends on depth of organic matter, seasonalwater table depths, fire and mineral input if any (Schafale and Weakly, 1990). Soil seriesassociated with theses plant communities (Croatan, Pamlico, Ponzer, Lynn Haven,Torhunta, Rutlege, and Pantego) are very strongly acidic (4.5-5.0) to ultra acidic (<3.5).Cation exchange capacity tends to be very low ranging from 1 to 30 cmol kg-1, but can beas high as 100 cmol kg-1 in the surface layers. Exchangeable phosphorus has found to bestrongly limiting to plant production in High Pocosin plant communities (Schafale andWeakly, 1990).

The North Carolina Department of Transportation (NCDOT) purchased a 256 hadrained Carolina bay in 1999, near Lumberton, North Carolina. This land was purchasedwith the intent of earning wetland mitigation credits. The Carolina bay being restored bythe NCDOT is called Juniper Bay. Juniper Bay was drained, cleared, and put intoagricultural production incrementally in 1971, 1981 and 1986. A review of the history ofthe clearing process was described by Ewing et al. (2003). Agricultural practices in theorganic soils include the addition of lime to achieve a pH of 5.5 to 5.0 (Lilly and Baird,1993). To reach this target pH up to 13470 kg ha-1 may have to be applied to overcomethe large reserve of acidic cations present in organic soils (Lilly, 1981). The NorthCarolina Cooperative Extension Service recommends the application of 135-180 kg N ha-

1, 35-55 kg P2O5 ha-1, and 90-115 kg K ha-1 for corn production (Crozer, 2000). Organicsoils in North Carolina have been shown to leach nitrogen and phosphorus and have to beapplied yearly (Lilly, 1981). Crop cultivation, depending on the crop, can include the useof chisel plows, moldboard plows, disks, and sub-soilers. The overall objective of theJuniper bay project is to restore Juniper bay back to an ecosystem that is typical ofCarolina bays. This includes restoration of hydrology, vegetation, and soils typicallyfound in Carolina bays.

The objective of this investigation was to describe and compare the physical andmorphological properties found in a natural Carolina bay, and those found in Juniper Bayand discuss the impacts on restoration efforts.

MATERIALS AND METHODS

Juniper Bay (JB) is located approximately 10 km southeast of Lumberton, NorthCarolina in Robeson County ((34o30’30”N 79o01’30”E). Juniper Bay was logged,drained and put into agricultural production in three stages. The western third of JuniperBay was drained in 1971, the central third and most of the eastern third was drained in1979, the area to the north in the eastern part was drained in 1986 (Figure 6.1). There isa perimeter ditch around the whole bay, two main drainage ditches that run NE/SW andmany lateral ditches that run NW/SE. The perimeter ditch and the main ditches areapproximately 7 m across and 4 m deep. The lateral ditches are roughly 1.5 m wide and1m deep. Areas that are enclosed by ditches are called ‘cuts’. There is only one outlet inJuniper Bay and it is located at the SW end of the main ditch on the western side. Initialsurvey of Juniper Bay indicated three types of soil based on the thickness of the organiclayer. The soils were classified as Aquic Haploarthods (<20cm organic material), Histic

128

Humaquepts (20-40 cm organic material), and Terric Haplosaprist (>40 cm organicmaterial), and will be referred to as mineral, histic and organic (Fig 1).

Sampling locations were chosen by randomly placing an equilateral triangle gridover the soils map. There were a total of 24 locations inside of Juniper Bay, 10 inorganic soil, 8 in mineral soil, and 6 in histic soils. Those points were then found in thefield using GPS. Using the point from the grid as a reference, a 1-1.5 m pit was dug atthe center of cut and near the closest lateral ditch, resulting in a pair of pits at eachlocation. Figure 2 shows the location of all pits. Soil profiles were described andsampled. Grab samples and Uhland cores (75mm in diameter x 75mm in height) forlaboratory analysis were taken from each horizon.

Three reference bays (RB) were selected in Bladen County, North Carolina basedon their similarity of soils to Juniper Bay, lack of drainage or agricultural applications,and ownership cooperation. The bays selected were Charlie Long-Millpond Bay(34o46’00”N 78o33’30”E) and Tatum Millpond Bay (34o43’00”N 78o33’00”E), bothlocated in the Bladen Lakes State Park and Causeway Bay (34o39’45”N 78o25’45”E)which is located 10 km N of White Lake, N.C. A trail from the rim of each bay to thecenter was cut through the vegetation (Fig 3). The general soil types found in JuniperBay were also found in all three natural bays, mineral, histic, and organic. Plots weremarked off and numbered at 50m intervals along the transects. Plots were then randomlyselected in each soil to be sampled. There were a total of eight sampling plots in eachreference bay, two in both the mineral and histic soils and four in the organic soil. Thesoil profile in the selected plot was described using a McCauley peat sampler and an openbucket auger. Grab samples were gathered and placed in plastic bags for laboratoryanalysis. The McCauley auger was also used to take undisturbed samples to determinebulk density and porosity (Soil Survey, 1993). Due to the high water table (10 cm belowthe surface) we were unable to dig a pit or obtain Uhland core samples. Samplingoccurred during the summer of 2002.

Grab samples from Juniper Bay and the reference bays were dried and groundwith an electric grinder to pass through a 2mm mesh sieve. Percent organic carbon andtotal nitrogen were determined through dry combustion with a Perkin-Elmer PE2400CHN Elemental Analyzer (Culmo, 1988). Uhland cores from Juniper bay were used todetermine saturated hydraulic conductivity, soil moisture release curves, bulk density,and porosity. Ten-cm undisturbed samples taken with the McCauley sampler from thereference bays were used to determine porosity and bulk density. Disturbed samplesfrom all horizons were placed in pressure plates to determine water content at 1 and 15bars in samples from Juniper bay, and 1/3 and 15 bars in samples from the referencebays. Available water was the difference in water content at 1/3 and 15 bars and porositywere calculated from the difference between saturation and oven dry.

Soil types, organic, histic, and mineral, were analyzed separately using the SASprocedure PROC MIXED (SAS, 2000), with an AR(1) covariance structure. There weretwo locations in the organic soil at Juniper Bay that had been drained 15 and 20 years,and 4 locations that had been drained for 30 years. There were three locations in the

129

histic soil at Juniper Bay that had been drained 15 and 20 years. There were fivelocations in the mineral soil at Juniper Bay that had been drained 20 years, and fourlocations that had been drained for 30 years. Each location had paired pits with onelocated at the crest of the field and one near the ditch. Horizons varied depending onlocation and pit and were evaluated as a spatially repeated measure. The data were notbalanced so LSMEANS were used to obtain estimated means. Comparisons among crestand near ditch pits were made with p>0.1 being significant.

The comparisons between the reference bays and Juniper bay were analyzedseparately depending on, soil type: (organic, histic, and mineral), using the SASprocedure PROC MIXED (SAS, 2000), with an AR(1) covariance structure. Data fromthe crest locations in Juniper Bay were used in this analysis after determining thatdifferences in crest and ditch locations were minimal. Organic, histic, and mineral soilswere combined across the three reference bays, resulting in 12 organic, six histic, and sixmineral locations. Horizons varied depending on location and pit and were evaluated as aspatially repeated measure. The reference Carolina bays were considered randomvariables so results could be inferred to other natural bays. Soils from the reference bayswere considered as time zero so comparisons could be made as to the effects of timesince drainage. The data were not balanced so LSMEANS were used to obtain estimatedmeans. Comparisons among Juniper Bay and the reference bays and time since drainagewere made with p>0.1 being significant.

RESULTS AND DISCUSSION

ORGANIC SOILS

Juniper vs. Reference Bays

A typical description of organic soil profiles at the reference bays and a typicalprofile or organic soil in Juniper Bay are given in Table 1. Reference bay soils includedOi and Oe horizons, which have been recognized in other Carolina bays with naturalvegetation (Stager and Cahoon, 1987). The Oi horizon is a fibric woody peat layer about20 cm thick that is composed of plant debris and roots. The Oe horizon is a hemic ormucky peat layer that is massive and often contains organic bodies and extended to50cm. These two layers are not present in JB and were probably removed and burnedduring clearing of the land for agriculture. The organic soil at JB has a surface horizonOap that is sapric material or muck that has a strong to moderate fine granular tosubangular blocky structure. The Oa horizons in the RB are composed of sapric materialwith massive structure and can extend from 40 to 180cm deep. The Oa horizons in JBhave sub-angular blocky to prismatic, and occasionally massive structure. Structuraldifference between JB and RB is a result of drainage and tillage. Similar changes instructure were described by Lee and Manoch (1974) in which organic soils becomegranular or subangular blocky in structure in the surface horizons; subsurface horizonsbecome prismatic with secondary blocky characteristics following drainage andpedogenic structure ends where the soil becomes saturated. This structure change shouldincrease saturated hydraulic conductivity. Oa horizons extend to depths of 40 and 70 cm

130

below the surface, much shallower than comparable locations in the RB. The thinnerorganic horizons in JB is a result of clearing activities and subsidence which, has beenestimated to be between 75 and 86 cm in JB (Ewing et al., 2003). Reported thickness ofthe organic layers in Carolina bays vary from 1cm (Newman and Schalles, 1990) to 4.6 m(Shariz and Gibbons, 1982). There is a thick (25cm) transitional horizon from organicmaterial to mineral material at the RB and the C horizons tend to be sandy loam or sand,while at JB the transition is a clear boundary and the C horizons are loamy sand to sandyclay loam. Saunders and Brown (1992) stated that the subsoil type in Carolina bays arerelated to their geomorphic surfaces, and could account for the differences in the Chorizon material.

The colors of the organic soils at Juniper bay were black to dark brown, (10YR)with an occasional redder hue (7.5YR) but were black at the RB (N 2.5/0). This issomewhat contradictory to what Dolman and Buol (1967) found in organic soils of theTidewater region of North Carolina and in which, deeper less oxidized organic horizonstended to be a dark reddish brown (5YR 2/2) while shallower more oxidized organichorizons tended to be black (10YR 2/1) to very dark brown (10YR 2/2). Lee andManoch (1974) found in a Wisconsin marsh in which there was a change in color from adark reddish brown to a black color after 50 years of drainage. However, the differencein color maybe a result of different soil environments. Lilly (1981) indicates that the darkreddish color of deep organic soils of North Carolina are preserved by being keptcontinuously anaerobic, while the areas with black organic soils have experienced someoxidization. This suggests that deeper horizons in Juniper Bay’s organic soils remainedanaerobic for longer periods and the RB formed under anaerobic conditions thatfluctuated.

Charcoal fragments were found in several of the profiles in JB, but were not seenin the RB suggesting that fire may have played a role in the development of the soils atJB. However, the profile sites described in the RB were limited in size and may not haveprovided adequate information for such a discovery. There have been other studies ofCarolina bays that have found evidence of fire in the soil profiles (Cohen et al., 1999).

LSMEANS estimates comparing physical characteristics of organic soils atJuniper Bay to the reference bays are in Table 2. There was a significant differencebetween Juniper Bay and the reference bays in available water. In the surface 20 cm ofJB and RB there was the same amount of available water, 0.27 and 0.30 cm3 cm-3.Available water in RB remained constant through the profile at 0.30 cm3 cm-3. Availablewater in JB increases to 0.50 cm3 cm-3 at 54 cm and then decreased back to levels foundin RB at 87 cm. Histosols have an extremely high water holding capacity and much of itis in the larger pores (gravitational) or in micropores that is unavailable for plant growth(Boelter and Blake, 1964). Fibric soils, like those in the RB, have large pores andrelatively high saturated hydraulic conductivities, while sapric soils have moremicropores and have hydraulic conductivities lower than clay-textured sediment (Boelter,1965). Despite the ability to hold large amounts of water, drained organic soils may dryout quickly and be prone to drought (Lilly, 1981). In addition, water in the small pores inthe underlying undeveloped organic soil will not move by capillary conductivity to the

131

larger pores in the developed surface soil, i.e. no recharge from underlying soil resultingin droughty conditions near the surface (Lilly, 1981). Although not significant, therewere trends in particle size, bulk density and porosity. There tended to be less sand andclay through the profile in the RB. These trends are probably a result of the geomorphicsurface (Saunders and Brown, 1992) or deposition from the surrounding landscape(Preston and Brown, 1964). Bulk density tended to be higher at JB and ranged from 0.48to 1.23 g cm-3 compared to 0.29 to 0.96 g cm-3 at the RB. Conversely porosity tended tobe higher at JB.

It is unfortunate that we were unable to obtain saturated hydraulic conductivitymeasurements in the RB. We would of expected higher Ksat values in the Oi and Oehorizons and lower Ksat values in the Oa horizons in the RB than at JB. Unrippenedorganic material K values of 0.002 to 0.19 cm hr-1 and 15-37cm hr-1 for a ripened to layer(Skaggs, 1976). Chaston and Siegel (1986) found that the K of decomposed peat canrange from 0.1 to 8.0 m day-1, with the large K values caused by discontinuous zones ofburied wood and other structural features which can enhance or obstruct water flow. Inthe upper layers of undisturbed peat K can be as high as 30m d-1 (Ingram 1967). The Kof geologic materials of low permeability is scale dependent: the greater the size of theflow system, the greater the permeability (Neuzil, 1986).

Time of drainage

There were significant differences in bulk density due to time of drainage in theorganic soils (Figure 4). Bulk density in organic soils tends to increase with subsidenceand decomposition (Boelter, 1965) and values will vary according to the amount ofmineral material and the type of vegetation. Bulk density also increases as traffic ofagricultural equipment compacts soil layers near the surface. Bulk density of JB organicsoils were consistently higher than RB organic soils and are similar to findings by Ewinget al., (2003b) in which bulk density in a drained Carolina bay is higher than that of anundrained Carolina bay. Bulk density at the surface of RB, was lowest at the surface andincreased with depth. Bulk density in the upper 50 cm was similar for all lengths ofdrainage JB to a depth of 100cm. Below 100cm bulk density was different among timeof drainage due to different types of mineral material in the C horizons. Bulk density washigher in the surface 20cm of JB and decreased at 50cm depth, as result of mineralmaterial from ditch spoil, subsidence, and tillage. The increase in bulk density below 50cm in JB is due to increases of mineral material as the soil changes from organic tomineral. Although not significant porosity tended to be the inverse of bulk densitythroughout the profile, when bulk density increased porosity decreased, and vise versa.

Crest vs. Ditch

Profiles of organic soils located at the crest of the field and near the ditch varieddue to increased drainage and addition of ditch spoil. Figure 5 provides an illustration ofthe soil at the crest and ditch at location 16, and Table 3 includes the associated profiledescriptions. The Oap horizon at the ditch is 24 cm deep and only 11cm deep at the crest,and both have the same strong medium granular structure. However, in the OA1 and

132

OA2 horizons the structure is massive at the crest while at the ditch there is weak tomoderate very coarse prismatic to sub-angular blocky structure, which indicates greaterprofile development. Both profiles have a dense or cemented layer between 100 and110cm. This is a characteristic of Carolina bays that has been described by others(Thom, 1970, and Wright et al., 2000), and indicates an accumulation of silica or humatematerial. There is a 2Bh horizon at the ditch that is not present at the crest. This maybe aresult of the increased movement water, and associated organic material, through the soilaround the ditch with the organic material accumulating in this horizon.

LSMEANS estimates comparing the crest and ditch locations in the organic soilscan be found in table 6.4. There was a significant difference in bulk density and totalpores between the crest and ditch in the organic soils. Bulk density near the ditch was0.16 to 0.46 g cm-1 higher in the upper 70 cm, and can be attributed to the higheramounts of sand deposited from maintaining the ditches. Bulk density for both areasdecreased from the surface to 50 cm, and then increased. Porosity was higher throughoutthe profile at the crest compared to the ditch as a result of the lower bulk density. Particlesize, saturated hydraulic conductivity, and available water were not found to besignificantly different between the crest and ditch location.

HISTIC SOILS

Juniper Bay vs. Reference Bays

A typical description of histic soil profiles at the RB and a typical profile or histicsoil in JB is given in Table 5. The histic soils tend to form a ring around areas of organicsoils in both JB and RB. The RB usually has an Oi or Oe horizon of woody peat thatconsists of accumulated plant debris and roots, with an Oa horizon of massive sapricmuck extending to less than 40 cm depth. Below this is an OC horizon where the organicmaterial changes to mineral material, and then a C horizon that is usually sand. There areno Oi or Oe horizons in the histic soils at JB. They have Ap horizons with organicallycoated sand grains, with an underlying Oa horizon that extends to approximately 40cmand tends to have granular to sub-angular blocky structure. The structure in the Oahorizon is due to agricultural practices and subsidence processes. There is a cleartransition in JB to mineral material that is usually sand or loamy sand. The color of the Chorizon in JB tends to be lighter in color (10YR 6/4) than in the RB (10YR 2/2), and is aresult of local geomorphic surfaces (Saunders and Brown, 1992).

LSMEANS estimates comparing physical properties of the histic soils at JuniperBay and the reference bays are in Table 6. There was a significant difference in availablewater and clay between JB and the RB. Juniper Bay tended to have 0.1to 0.3 cm3 cm-3

more available water than the RB. The higher amount of clay in JB, approximately 8 to25% higher through the profile, can result in higher amounts of available water. Inaddition, there is the potential that the organic horizons at the RB may contain a greaterportion of macropores reducing the amount of available water. There were nosignificant difference between JB and the RB in percent sand, bulk density and porosity.

133

Time since drainage

There was a significant time since drainage effect on particle size and availablewater (Fig. 6). Sand ranged from 72.9% at 32 cm and increased with depth to 92.9% at119cm at time zero. After 15 years sand ranged from 78.7 to 63.4% and 47.2 to 94.2%after 20 years. Clay was the lowest at time zero ranging from 7.2 to 2.8%. After 15 yearsof drainage clay ranged from 30.5% in the C horizon to 21.2% at the surface, and after 20years it ranged from 3.6 to 42.4%. Available water was similar in the upper 20 cm for 0and 15 years since drainage, 0.37 and 0.33 cm3 cm-3 respectively, but lower 20 yearssince drainage, 0.21 cm3 cm-3. All increased as depth increased to 32cm, however 0 and20 years were similar, 0.42 and 0.43 cm3 cm-3 respectively and 15 year was higher at 0.65cm3 cm-3. Below 32cm, 0 year decreased and remained around 0.17 cm3 cm-3 through therest of the profile. The 15 years since drainage decreased to 0.11 cm3 cm-3 by 85cm andthin increased to 0.22 cm3 cm-3 at 104cm. The area drained for 20 years increased to 0.54cm3 cm-at 77cm, then decreased to 0.25 cm3 cm-3 at 122cm.

Crest vs. Ditch

There are some visual differences between soil profiles located at the crest and theditch through the histic soils. Figure 7 provides an illustration of the soil at the crest andditch at location 11, and table 6.7 is the associated profile descriptions. There was anOap horizon at the surface of the crest location with strong medium granular structure,while the ditch location had an Ap horizon with a weak fine granular structure. Thissuggests that either the organic soil has decomposed to a point that it is no longer organic,or that mineral material from the ditch has been mixed in. The OA horizon at the ditchextends to a depth of 23cm, while at the crest it extends to a depth of 34cm, againindicating a loss of organic material, from oxidization or land shaping. The more definedstructure in the OA horizon at the ditch, weak medium sub-angular blocky, compared tothe weak coarse prismatic at the crest, indicates drier conditions which results in theformation of better structure in organic soils (Lee and Manoch, 1974). There are somelocations in the histic soils where it is histic at the center of the ditch, but is closer to“mineral” near the ditch. Also there are locations near the ditch in which the surfacehorizon has been influenced by spoil form the ditch maintenance. All organic soils inthe, if cultivated long enough, will become mineral soils with dark surface horizons dueto the inevitable natural loss of organic matter through oxidization (Lilly, 1981), and thismay be the case in the histic soils of Juniper Bay.

There were no physical differences between the crest and ditch positions in thehistic soils, however, LSMEANS estimates for the physical properties of the histic soilcan be found in table 6.8. However, saturated hydraulic conductivity tended to be higherin soils at the ditch, probably due to increased soil structure.

134

MINERAL SOILS

Juniper Bay vs. Reference Bay

Mineral soils in RB and JB were relatively similar (Table 9) with differences inthe surface horizon being a result of agricultural activities. The mineral soils at the RBhad a thin (10cm) Oi horizon consisting of plant debris and roots, which are not presentin JB. Below this is an A horizon with weak granular structure. A thick (20cm) Ehorizon is next, followed by several spodic horizons. There were a few locations in theRB that had an E’ and Bh’ horizons. The mineral soils in JB had a dark Ap horizon withorganically coated sand grains and weak granular structure. There was a thick (20 cm) Ehorizon, which sometimes was mixed with the Ap horizon, followed by several Bhhorizons. There were also C horizons at the bottom of some mineral soil profiles thatconsisted of sand or cemented sandy clay loam similar to what was described by Ingramet al., (1959), Stager and Cahoon (1987), and Lide et al., (1995).

There were no significant differences in particle size between the mineral soils atJuniper Bay and reference bays (Table 10). We were unable to obtain undisturbedsamples from the RB to determine bulk density, porosity and available water. Howeverbulk density at JB ranged from 1.09 g cm-3 at the surface and increased to 1.63 g cm-3 at126cm, porosity ranged from 0.42 to 0.52 cm3 cm-3, and available water ranged from0.27 to 0.18 cm3 cm-3.

Time since drainage

We were unable to take intact samples in the RB to determine bulk density,porosity, and available water. The areas drained for 30 years had an available water of0.23 in the surface 10cm, increased to 0.35 at 23cm, then decreased to 0.18 at 57cm, andthen remained close to the values found in the 20 year areas (Fig. 8). Bulk densities inmineral soils at JB were similar after 20 and 30 years of drainage.

Crest vs. Ditch

There are visual differences between soil profiles located at the crest and theditch. Figure 9 provides an illustration of the soil at the two locations, and Table 11 is theassociated profile description of location 3. The Ap horizon is approximately the samethickness and color at both locations, however, there is 10% less organically coated sandgrains at the ditch. It appeared that some of the surface material might be from ditchmaintenance, and is more evident at other locations. There is an AE horizon at the ditchthat is a result of tillage. Such a mixing of A and E horizons occurred at several locationsat both the crest and ditch in the mineral soils. The E horizon is 4 cm thicker at the ditch,and the Bhir horizons extend deeper into the profile. There was defined structure in thesurface 22cm at the ditch but only the surface 12 cm at the crest. There was a sulfursmell associated with the lowest horizon indicating reduced conditions.

135

There was a significant difference in the particle size between the crest and ditchpositions in the mineral soil (Table 12). There was more than 85% sand throughout theprofile in both positions. The crest had 3% more sand in the surface 15 cm,approximately the same at 30 cm and the crest had 7.5% more at 50 cm, andapproximately the same for the remainder of the profile, increasing with depth to 95%sand at 125cm. There was 2 to 6% more clay in the surface 50 cm at the ditch positionthan the crest. Below 50 cm clay content was similar, approximately 3%, for the rest ofthe profile. There were no significant differences between the crest and ditch for bulkdensity, porosity, available water, or saturated hydraulic conductivity (Table 12).However , there was a trend for saturated hydraulic conductivity to be higher near theditch and for available water to be lower. This trend probable reflects the increaseddevelopment of structure and the associated increase in macropores.

CONCLUSIONS

The organic soils in Juniper bay have under gone the most change since drainagebegan. There were some differences between crest and ditch locations, however they arerelatively small compared to the differences between JB and the RB. Agriculturalproduction caused the removal of surface organic horizons Oi and Oe. The remainingorganic soil is thinner due to subsidence and has developed granular structure at thesurface due to tillage and desiccation. The loss of the surface horizons and thedevelopment of structure have increased hydraulic conductivity, and plant availablewater. Bulk density has increased since drainage and indicated that the process ofoxidation is still ongoing. Restoration efforts may be hampered or influenced by thechange in hydraulic properties. For example restoring the water table to pre-drainagelevels might result in a water table that is above the soil surface, or the sapric Oa horizon,which perched water originally due to extremely low Ksat, may allow water to movethrough the profile. Although there were differences in particle size properties in theorganic soils, we feel that there is not a large enough difference to influence restorationefforts. The dense layer found beneath the organic soils in Juniper Bay was notencountered in the reference bays. This layer was either deeper in the profile at thereference bays or has formed since drainage at Juniper Bay and is now acting as anaquitard.

Histic soils at Juniper bay were lacking an Oi and Oe horizons and have greaterstructure development. Other physical properties were similar enough between JB andthe RB and should not affect hydraulic properties. However, some histic areas in JuniperBay show indications that they may have once been organic soils and have subsided toonly a histic epipedon. This may result in water tables above the soil surface ifhydrology is returned to the pre-drained levels. The effects of ditch maintenance weremost evident in the histic soils with an Ap horizon over an Oa horizon at the ditchlocations. There were also indications of better structure at the ditch indicating betterdrainage. Difference between the ditch and crest and differences in particle size data arenot enough to justify different management practices.

136

Mineral soils at all locations were well developed with some differences. Thesedifferences include a thin Oi horizon at the RB and a thicker Ap horizon and deeper Ehorizon as a result of the addition of ditch spoil. The E horizon was thicker at the ditchand the Bh horizons extended deeper into the profile indicating better water movementthrough the profile. Soil structure was more developed at Juniper Bay. There is little toindicate that the physical properties of the mineral soils in JB are different than those ofthe RB, and both should behave similarly under the same hydraulic conditions.Differences in sand silt and clay are relatively small and could be attributed to thedepositional environments at the time of formation. Areas of JB that are now mineralsoils may have had histic epipedons prior to drainage and may result in water tablesabove the soil surface if hydrology is returned to the pre-drained levels.

The restoration of Juniper Bay soils to those typically found in natural undrainedCarolina bay wetlands can be achieved over time. The mineral soils are similar enoughalready, however, the mineral soils may not support hydrophytic vegetation as well as thehistic and mineral soils. Histic soils would require some time to accumulated organicdebris to form an Oi and Oe horizons. The organic soils would require a great deal oftime to be restored to natural conditions, due to the large amounts of organic material thathas been lost since drainage.

137

REFERENCES

Bliley, D.J. and D.A. Burney. 1988. Late Pleistocene climatic factors in the genesis of aCarolina Bay. Southeastern Geo. 29:83-101.

Boelter, D.H. 1965. Hydraulic conductivity of peats. Soil Sci. 100:227-231.

Boelter, D.H. and G.R. Blake. 1964. Importance of volumetric expression of watercontents of organic soils Soil Sci.Soc Am . Proc. 28:176-178.

Chason, D.B. and D.I. Siegel. 1986. Hydraulic conductivity and related physicalproperties of peat, Lost River peatland, northern Minnesota. Soil Science 42:91-99.

Cohen, A.D., C.P.Gage, W.S. Moore, and R.S. VanPelt. 1999. Compining organicpetography and palynology to assess anthropogenic impacts on peatlands Part 2.An example from a Carolina Bay wetland at the Savannah River Site in SouthCarolina. Int. J. Coal Geo. 39:47-95.

Crozer, C. 2000. Fertilizer and lime management. North Carolina Corn ProductionGuide. North Carolina Cooperative Extension Service, North Carolina StateUniversity, Raleigh, N.C.

Culmo, R.F. 1988. Principle of Operation – The Perkin-Elmer PE 2400 CHN ElementalAnalyzer. Perkin Elmer Corp. Norwalk, CT.

Ewing, J.M. 2003a. Subsidence chapter Dissertation.

Ewing, J.M, M.J. Vepraskas, and C.W. Zanner. 2003b Agricultural Impacts on Soils in aDrained Carolina Bay. North Carolina Society of Soil Scientists, 2003.

Erwin, K.L. 1991. An Evaluation of Wetland Mitigation in the Southeast. In J.A. Kuslerand M.E. Kentula eds. Wetland Creation and Restoration. Island Press,Washington DC, pp. 233-265.

Gallihugh, J.L. and J.D. Rogner. 1998. Wetland Mitigation and 404 Permit ComplianceStudy, Vol. I. U.S. Fish and Wildlife Service, Region III, Burlington, IL, and U.S.Environmental Protection Agency, Region V, Chicago. 161pp.

Grant C. 1945. A biological explanation of the Carolina Bays. The Scientific Monthly.61:443-450.

Ingram, H.A.P. 1967. Problems of hydrology and plant distributions in mires. J.Ecology 55:711-724.

Ingram, R. L. and M. Robinson. 1957. Clay minerals of some Carolina Bay sediments.J. Mitchell Soc. N. C. Acad. of Sci. Proc. 73:241.

138

Johnson, D.W. 1942. Origin of the Carolina Bays. Columbia University Press.

Johnson. D.W. 1936. Origin of the Supposed Meteorite Scares of the Carolina. Science. 84:15-18.

Lee, G.B. and B. Manoch. 1974. Macromorphology and Micromorphology of aWisconsin Saprist. In Histosols Their Characteristics, Classification and Use.M. Stelly and R.C. Dinauer ed. SSSA pub 6. Soil Sci. Soc. Am. Madison,Wisconsin.

LeGrand, H. E. 1953. Streamlining of the Carolina Bays. J. Geol. 61:263-274.

Lide, R. F., V.G. Meentemeyer, J. E. Pinder III, and L.M. Beatty. 1995. Hydrology of aCarolina Bay located on the upper coastal plain of Western South Carolina.Wetlands. 15:47-57.

Lilly, J.P., 1981. The Blackland Soils of North Carolina: Their Characteristics andManagement for Agriculture. North Carolina Ag. Res. Tech Bul. No. 270. pp. 27-33.

Lilly, J.P. and Baird. 1993. North Carolina Cooperative Extension Service Pub. AG439-17. North Carolina State University: Raleigh, N.C.

Mehlich A. 1984. Mehlich-3 soil test extractant: a modification of Mehlich-2 extractant.Commun. Soil Sci. Plant Anal. 15(12):1409–16.

Mehlich A. 1976. New buffer method for rapid estimation of exchangeable acidity andlime requirement. Commun. Soil Sci. Plant Anal. 7(7): 637–52.

Neuzil, C. 1986. Groundwater flow in low-permeability environments. Water ResourcesResearch. 22:1163-1197.

Nifong, T.D. 1998. An ecosystem analysis of Carolina Bays in the Coastal Plain of theCarolinas. Ph.D. Dissertation. University of North Carolina, Chapel Hill, NorthCarolina.

Odum, H. T. 1952. The Carolina Bays and a Pleistocene weather map. Am. J. Sci.250:263-270.

Preston, C.D. and C.Q. Brown. 1964. Geologic section along a Carolina Bay, SumterCounty, South Carolina. Southeastern Geology. 6:21-29.

Prouty, W.F. 1952. Carolina Bays and their Origin. Bulletin of the Geological Societyof America. 63:167-224.

139

SAS Institute Inc. 2000. SAS version 8.2. Cary, NC, USA.

Saunders, C.L. and C.Q. Brown. 1992. Substrate variability and internal sediments andsediment textural gradients in three Carolina bays, south-central coastal plain,

North Carolina. 41st Southeast GSA meetings abstracts. 24:62-63.

Skaggs, R.W. and J.S. Barnes. 1976. Drainage and water table control on colloidal mucksoils. Am .Soc. Agric. Eng. Annual meeting, Paper No. 76-2041.

Soil Survey Division Staff, 1993. Soil Survey Manual. United States Department ofAgriculture, Washington D.C.

Stager, J.L. and L.B. Cahoon. 1987. The age and tropic history of Lake Waccamaw,North Carolina. J. Elisha Mitchell Sci. Soc. 103:1-13.

Stolt, M. H. and M.C. Rabenhorst. 1987. Carolina Bays on the Eastern shore ofMaryland I. Soil characterization and classification. Soil Sci. Soc. Am. J. 51:394-398.

Thom, B.G. 1970. Carolina Bays in Horry and Marion Counties, South Carolina. Geo.Soc. Am. Bull. 81:783-814.

US Army Corps of Engineers. 2002. Regulatory Guidance Letter No. 02-2.

Weakley, A.S. and S.K. Scott. 1982. Natural features summary and preserve design forCarolina Bays in Bladen and Cumberland Counties, NC. Unpublished report toNC Natural Heritage Program, The Nature Conservancy, and Earthlines, Inc.,Raleigh,NC.

Wilson, K.A. and W.J. Mitsch. 1996. Functional assessment of five wetlands constructedto mitigate wetland loss in Ohio, USA. Wetlands 16:436-451.

Wright, E., T. Carter, R. Leeson, S. Forman, L. Abrams, and S. Harris. 2000.Stratigraphy of two-overlapping Carolina Bays, Horry County, South Carolina.GSA proc. 32:84.

140

Table 1. Example profile descriptions of organic soils from the reference bays andJuniper Bay.

Horizon Depth Description

Reference bay

Oi 0-20 Black (10YR 2/1) fibric to hemic muck that is part of the root mat;

weak coarse platy structure from layering of plant debris; gradual

boundary.

Oe 20-50 Very dark brown (10YR 2/2) hemic muck; massive structure with

many organic bodies 0.5 to 1.0 cm in diameter; gradual boundary.

Oa 50-85 Black (10YR 2/1) sapric muck; massive structure; gradual

boundary.

OC 85-110 Very dark brown (10YR 2/2) mucky loam to mucky sandy loam;

massive structure; gradual boundary.

2C1 110-140 Very dark brown (10YR 2/2) sandy loam; massive structure;

gradual boundary.

2C2 140-180 Dark gray (10YR 4/1) sand; single grain structure.

Juniper Bay: JB10C

Ap 0-10 Very dark brown (10YR 2/2) sandy loam; moderate medium (3mm)

granular structure; abrupt boundary.

Oa 10-20 Black (N 2.5/0) sapric material; moderate fine (2mm) granular to

sub-angular blocky structure; abrupt boundary.

OA1 20-51 Black (10YR 2/1) mucky silt loam, weak very coarse (20cm)

prismatic structure; clear boundary.

141

OA2 51-61 Very dark brown (10YR 2/2) mucky silt loam with 25% black (N

2.5/0) charcoal; weak coarse (10cm) prismatic structure.

OA3 61-68 Dark brown (10YR 3/3) mucky fine sandy loam; very weak coarse

(10cm) prismatic structure; clear boundary.

Bw 68-107 Yellowish brown (10YR 5/4) very fine sandy clay loam; weak very

coarse (20cm) platy to moderate medium (5cm) sub-angular blocky

structure; clear boundary.

C 107-120 Light brownish yellow (10YR 6/2) sandy clay loam; strong medium

(1cm) angular blocky structure; faint reaction to alpha-alpha.

142

Table 2. Comparison of organic soils between Juniper Bay (JB) and Reference bays RB.

Horizont Depth Sand Clay Bulk Density PorosityAvailableWater*

(cm) -------- (%) ------- (g cm-3) (cm cm-3) (cm cm-3)

Juniper Bay

1 15 57.5 26.4 0.70 0.60 0.272 32 53.8 30.0 0.52 0.64 0.463 54 57.1 27.8 0.48 0.67 0.504 75 59.5 22.2 0.85 0.62 0.425 87 71.1 11.8 0.12 0.58 0.306 108 --- --- --- --- ---

Error 14.8 6.7 0.17 0.14 0.07

Reference Bays

1 19 --- --- --- --- 0.302 48 40.5 16.2 0.29 0.62 0.293 88 44.4 13.6 0.35 0.66 0.264 123 55.7 12.0 0.64 0.58 0.275 148 53.5 12.9 0.74 0.60 0.326 171 63.8 11.2 0.96 0.52 0.19

Error 10.9 4.5 0.15 0.09 0.05* significant at p>0.1 (---) data not available. t Horizons were different between Juniper

bay and reference bays and are numbered here to indicate the presence of adistinct horizon.

Table 3. Profile descriptions of crest and ditch in an organic soil at Juniper Bay.

143

Horizon Depth Matrix Texture Structure OC Comments(cm) (%)

JB16COap 0-11 N 2.5/0 sm Strong

MediumGranular

26.2

OA1 11-31 10YR2/2

sm Massive 52.5 10% wood fragments10% charcoal

OA2 31-52 10YR2/2

sm Massive 27.9 10% wood fragments10% charcoal

2Bw 52-61 5YR 3/3 SL ModerateCoarse

Prismatic

5.5

2BC 61-78 10YR5/4

LS Massive 0.5 10% wood fragments40% 2.5Y 8/1 depletions

2C1 78-100 10YR6/4

LS Massive 0.3 5% wood fragments

2C2 100+ 10YR4/1

LS Single Grain 0.3 Dense; reacts to alpha,alpha

JB16DOap 0-24 N2.5/0 sm Strong

ModerateGranular

13.4

OA1 24-41 10YR2/2

sm ModerateVery CoarseSubangular

Blocky

31.9 20% wood fragments

OA2 41-59 10YR2/1

sm Weak VeryCoarse

Prismatic

19.6 10% wood fragments

2Bw 59-74 10YR5/4

SL Very WeakCoarse

SubangularBlocky

1.5 20% 10YR 8/2depletions

2BC 74-110 10YR6/4

LS Massive 0.5 10% 10YR 8/2depletions, slightly brittle

2Bh 110-130

10YR2/1

SL Massive 0.5 Cemented, brittle andfirm

sm = sapric material, LS = loamy sand, SiL = Silt loam, SL = Sandy loam, mSL = muckysandy loam

144

Table 4. Crest vs. Ditch comparison of physical properties of the organic soil in JuniperBay.

Horizon Depth Sand ClayBulk

Density* Ksat Porosity*Available

Water

(cm) -------- (%) ------- (g cm-3) (cm hr-1) (cm cm-3) (cm cm-3)

Crest

Oap 15 56.4 27.3 0.69 10.35 0.61 0.17OA1 32 53.8 30.0 0.53 6.76 0.62 0.35OA2 52 57.1 27.7 0.47 5.41 0.67 0.41Bw 74 59.5 22.2 0.83 7.48 0.65 0.33BC 88 71.1 11.8 1.16 1.49 0.61 0.20C 110 --- --- --- --- --- ---

Error 3 13.8 6.5 0.10 3.20 0.04 0.07

Ditch

Oap 15 70.0 17.7 0.85 7.13 0.60 0.15OA1 32 66.7 18.5 0.69 10.17 0.65 0.26OA2 48 70.2 17.3 0.87 4.43 0.59 0.24Bw 66 72.2 13.7 1.29 2.62 0.46 0.34BC 92 72.4 14.2 1.23 3.22 0.47 0.20C 107 79.4 10.7 1.61 2.37 0.52 0.47

Error 3 13.8 6.5 0.10 3.20 0.04 0.07* significant at p>0.1 (---) data not available.

145

Table 5. Example profile descriptions of histic soils from the Reference bays andJuniper Bay.

Horizon Depth Description

(cm) Reference bay

Oi 0-15 Very dark brown (10YR 2/1) hemic material; weak medium platy

Structure; clear boundary; root mat.

Oa 15-36 Black (N 2.5/0) sapric material; massive structure; clear boundary.

OC 36-65 Black (N 2.5/0) sandy loam; massive structure; gradual boundary.

C 65-100 Very dark brown (10YR 2/2) sand; single grain structure.

Juniper Bay: JB61C

Ap 0-22 Black (N 2.5/0) mucky sandy loam with 60% organically coatedsand grains; moderate medium granular structure; clear boundary.

Oa 22-36 Very dark brown (10YR 2/2) sapric material; strong coarse granular

structure; clear boundary.

Bw 36-56 Dark brown (10YR 3/3) sandy loam with 2% gray (10YR 6/1) sand;

weak medium sub-angular blocky structure; clear boundary.

BC 56-74 Dark yellowish brown (10YR 4/6) loamy sand with 2% gray

(10YR 6/1) sand; weak medium sub-angular blocky structure;

gradual boundary.

C 74-108 Brownish yellow (10YR 6/6) with 2% gray (10YR 6/1) sand; single

grain structure.

146

Table 6. Comparison of histic soils between Juniper Bay (JB) and Reference bays RB.

Horizon t Depth Sand Clay*Bulk

Density PorosityAvailableWater*

(cm) -------- (%) ------- (g cm-3) (cm cm-3) (cm cm-3)

Juniper Bay

1 18 70.9 22.0 0.84 0.66 0.272 30 57.4 32.0 0.63 0.62 0.543 51 64.4 28.6 1.16 0.56 0.414 65 81.8 13.7 1.32 0.48 0.465 95 82.4 10.4 1.73 0.32 0.226 112 75.9 19.6 1.67 0.40 0.23

Error 4.6 8.5 5.6 0.07 0.05 0.13

Reference Bays

1 16 --- --- --- --- ---2 32 72.9 7.2 0.6 0.62 0.443 56 85.9 3.2 0.61 0.61 0.184 82 91.1 2.8 0.47 0.37 0.195 118 92.9 2.8 --- --- ---

Error 4.4 5.4 3.7 0.11 0.07 0.10* significant at p>0.1 (---) data not available.t Horizons were different between Juniper bay and reference bays and are numbered here

to indicate the presence of a distinct horizon.

Table 7. Profile descriptions of crest and ditch in a histic soil at Juniper Bay.

147

Horizon Depth Matrix Texture Structure OC Commentscm %

JB11COap 0-18 N 2.5/0 sm Strong

MediumGranular

46.4

OA 18-34 10YR 2/2 sm WeakCoarse

Prismatic

48.8 5% wood fragments

Bw 34-57 10YR 2/2 SiL WeakMediumPrismatic

6.1

2BC 57-67 10YR 6/4 LS SingleGrain

0.6 15% 10YR 8/1depletions

2C1 67-90 10YR 7/6 LS SingleGrain

0.1 20% 10YR 8/1depletions

2C2 90-110 10YR 6/4 LS SingleGrain

0.2

JB11DAp 0-10 N2.5/0 SL Weak Fine

Granular5.2 90% coatings

OA 10-23 N2.5/0 mSL WeakMedium

SubangularBlocky

12.6 80% coatings

Bw1 23-39 10YR 2/2 SL Weak FinePrismatic

2.7 20% 10YR 8/1depletions

Bw2 39-59 10YR 3/4 SL VeryWeak

MediumPrismatic

1.0 30% 10YR 8/2depletions

BC 59-92 10YR 6/6 LS SingleGrain

0.3

C1 92-111 10YR 5/2 LS Massive 0.3 1-2cm bands of10YR 3/1

C2 111-120 10YR 2/1 SL Massive 0.8 Sulfur smellsm = sapric material, LS = loamy sand, SiL = Silt loam, SL = Sandy loam, mSL = muckysandy loam

148

Table 8. Crest vs. Ditch comparison of physical properties of the histic soils in JuniperBay.

Horizon Depth Sand ClayBulk

Density Ksat PorosityAvailable

Water

(cm) -------- (%) ------- (g cm-3) (cm hr-1) (cm cm-3) (cm cm-3)

Crest

Oap 17 70.9 22.0 0.82 9.29 0.06 0.19OA 30 57.4 31.9 0.65 4.39 0.64 0.42Bw 52 64.4 28.6 1.15 3.22 0.56 0.34BC 70 81.8 13.7 1.36 2.23 0.47 0.36C1 97 82.4 10.4 1.67 2.88 0.35 0.16C2 111 75.9 19.6 1.70 1.33 0.34 0.13

Error 7.56 6.90 0.11 3.10 0.07 0.10

Ditch

Oap 15 74.8 19.5 0.89 17.16 0.36 0.14OA 30 66.9 22.1 0.65 10.56 0.73 0.16Bw 45 77.6 15.2 1.06 5.44 0.57 0.18BC 66 79.4 13.6 1.40 5.37 0.35 0.45C1 85 82.6 14.7 1.60 2.06 0.31 0.11C2 99 79.1 16.9 1.76 0.03 0.40 0.15

Error 7.56 6.90 0.11 2.8 0.07 0.10* significant at p>0.1

149

Table 9. Example profile descriptions of mineral soils in reference bays and JuniperBay.

Horizon Depth Description

(cm) Reference bay

Oi 0-10 Brown (7.5YR 3/4) fibric organic material, root mat and leaf litter;

weak medium platy structure; clear boundary.

A 10-20 Black (10YR 2/1) sandy loam; weak fine granular structure; clear

boundary.

E 20-40 Gray (10YR 6/1) sand; single grain structure; clear boundary.

Bh1 40-70 Black (10YR 2/1) sandy loam; weak medium sub-angular blocky

Structure; clear boundary.

Bh2 70-100 Very dark grayish brown (10YR 3/2) loamy sand; weak coarse

sub-angular blocky structure.

Juniper Bay: JB17C

Ap 0-20 Black (10YR 2/1) loamy sand with 95% organically coated sand

grains; weak medium (2-5mm) granular structure; abrupt

E 20-47 20-47cm. Gray (N 6/0) loamy sand with 12% A horizon material

in root channels; single grain structure; clear boundary.

Bhir1 47-70 Very dark brown (10YR 2/2) sandy loam with 25% E material in

upper half of horizon; massive structure; clear boundary.

Bhir2 70-92 Black (10YR 2/1) and dark brown (10YR 3/3) sandy loam in

alternating layers 2-7cm thick; massive structure; abrupt boundary.

Bh 92-120 Black (10YR 2/1) sandy loam with 10% gray (10YR 6/1) sandy

grains; massive structure; firm and slightly brittle.

150

Table 10. Comparison of mineral soils between Juniper Bay (JB) and Reference baysRB.

Horizon t Depth Sand ClayBulk

Density PorosityAvailableWater**

(cm) -------- (%) ------- (g cm-3) (cm cm-3) (cm cm-3)

Juniper Bay

Ap 10 90.9 7.4 1.09 0.48 0.23A/E 27 88.1 5.1 1.32 0.46 0.26E 44 93.8 3.1 1.45 0.49 0.26

Bhir1 62 93.2 3.7 1.49 0.52 0.20Bhir2 85 94.5 4.3 1.55 0.49 0.23

Bh 101 95.7 2.5 1.59 0.43 0.17C 126 95.9 2.7 1.63 0.43 0.18

Error 7 3.5 1.1 0.04 0.06 0.07

Reference Bays

Oi 10 82.7 4.1 --- --- ---A 22 85.7 3.3 --- --- ---E 52 96.4 1.2 --- --- ---Bh 82 90.4 4.2 --- --- ---

Error 6 2.6 0.8 --- --- ---* significant at p>0.1 (---) data not available. t Horizons were different between Juniperbay and reference bays and are numbered here to indicate the presence of a distincthorizon.

151

Table 11. Profile descriptions of crest and ditch in a mineral soil at Juniper Bay.

Horizon Depth Matrix Texture Structure OC Commentscm %

JB03CAp 0-12 N 2.5/0 LS Moderate

FineGranular

3.3 90% coated sands

E 12-26 5YR 7/1 LS SingleGrain

0.2 25% Ap materialmixed in

Bhir 26-42 5YR 3/1 LS Massive 1.1Bir 42-61 10YR 6/3 LS Single

Grain1.2

B’hir 61-75 10YR 5/2 LS SingleGrain

0.4 Accumulationorganic material

C 75-85+ 10YR5/2 LS SingleGrain

0.9 Sulfur smell

JB03DAp 0-15 N 2.5/0 LS Moderate

FineGranular

2.3 80% coatings

A/E 15-22 N 2.5/02.5Y 7/1

LS Weak FineSubangular

Blocky

1.7 Mixed Ap and Ehorizons

E 22-40 2.5Y 7/1 LS SingleGrain

0.1

Bhir1 40-54 10YR 3/2 LS Massive 0.8 2 cm bands of 10YR2/1

Bhir2 57-70 10YR 3/3 LS Massive 0.8Bhir3 70-84 10YR 3/3 LS Massive 1.5 25% 10YR 2/1

weakly cemented LSBhir4 84-100 10YR 2/2 LS Massive 2.1 Sulfur smell

LS = loamy sand

152

Table 12. Crest vs. Ditch comparison of physical properties of the mineral soils inJuniper Bay.

Horizon t Depth Sand* Clay*Bulk

Density Ksat PorosityAvailable

Water(cm) ------ (%) ------ (g cm-3) (cm hr-1) (cm cm-3) (cm cm-3)

Crest

1 15 90.8 7.45 1.09 18.4 0.05 0.132 31 88 5.21 1.31 15.38 0.46 0.163 48 93.7 3.21 1.44 9.97 0.48 0.164 67 93.1 3.81 1.48 5.42 0.53 0.105 92 94.4 4.31 1.54 5.14 0.49 0.126 107 95.3 2.99 1.58 5.64 0.43 0.087 122 95.5 3.19 1.61 3.32 0.43 0.08

Error 1.52 1.2 0.05 4.06 0.04 0.03

Ditch

1 15 87.7 10.24 1.19 19.01 0.45 0.142 32 89.1 8.13 1.27 26.12 0.43 0.113 49 85.6 9.34 1.47 21.52 0.35 0.084 70 90.7 3.79 1.51 6.71 0.43 0.125 89 93.1 4.28 1.49 6.65 0.44 0.136 107 94.0 4.18 1.52 4.27 0.39 0.137 125 95.1 3.14 1.54 2.7 0.38 0.14

Error 1.52 1.2 0.05 4.06 0.04 0.03* significant at p>0.1 t Horizons were different between Juniper bay and reference bays

and are numbered here to indicate the presence of a distinct horizon.

153

Figure 1. Aerial photo of Juniper Bay (2.4 x 1.6 km) showing the time of drainage andareas where organic, histic, and mineral soils are located.

154

Figure 2. Pit locations at Juniper Bay

Ditches

Soil Pit Locations

# Soilpits.shp# Justinjuly83m.shp

Ditches.shp

200 0 200 400 600 800 1000 Meters

N

##

##

##

##

##

##

##

##

##

##

##

##

##

##

##

#####

##

##

#

##

##

##

##

##

10

5

9

14 15

11

6

7

1312

17

16

1

8

3

63 2

4

66

61

62

64

67

6568

155

Figure 3. Location of transects in Causeway Bay, 1.8 x 1.15km (a), Charlie Long

Millpond Bay, 1.9 x 1.2 km (b), and Tatum Millpond Bay, 4.4 x 2.2 km(c).

a.

c.

b.

156

0

20

40

60

80

100

120

140

160

180

0.2 0.4 0.6 0.8 1

Total Pores cm3 cm-3

Dep

th (

cm)

0

15

20

30

a

0

20

40

60

80

100

120

140

160

180

0 0.5 1 1.5 2

Bulk Density(g cm-3)

Dep

th (

cm)

0

15

20

30

b.

Figure 4. Effects of drainage time on total pores (a) and bulk density (b) of an organicsoil. Standard deviation is ± 0.1 g cm-3.

157

Figure 5. Photo comparison of soils at the crest (left) and ditch (right) at location 16 in

in organic soil in Juniper Bay. Scale is in cm.

158

0

20

40

60

80

100

120

140

40 60 80 100

Sand(%)

Dep

th (

cm)

0

15

25

a

0

20

40

60

80

100

120

140

0 10 20 30 40 50

Clay(%)

Dep

th (

cm)

0

15

25

b

159

0

20

40

60

80

100

120

140

0 0.2 0.4 0.6 0.8

Avaiable Water(cm3 cm-3)

Dep

th (

cm)

0

15

20

c

Figure 6. Effects of drainage time on sand (a), clay (b), and available water (c) in ahistic soil.

160

Figure 7. Photo comparison of soils at the crest (right) and ditch (left) at location 11 in a

soil with a histic epipedon in Juniper Bay. Scale is in cm.

161

0

20

40

60

80

100

120

140

0 0.1 0.2 0.3 0.4

Available Water(cm3 cm-3)

Dep

th (

cm)

0

20

30

a

0

20

40

60

80

100

120

140

1 1.2 1.4 1.6 1.8

Bulk Density(g cm-3)

Dep

th (

cm)

0

20

30

bFigure 8. Effects of drainage time on available water (a) and bulk density (b) in amineral soil.

162

Figure 9. Photo comparison of soils at the crest (left) and ditch (right) at location 3 in a

mineral soil in Juniper Bay

163

Chapter 7

ESTIMATING PRIMARY AND SECONDARY SUBSIDENCE IN AN ORGANIC SOIL15, 20, and 30 YEARS AFTER DRAINAGE

J.M. Ewing and M.J. Vepraskas

INTRODUCTION

Organic soils form by the accumulation of plant debris under anaerobic conditions (Buolet al. 2003). Most organic soils occur in areas that are saturated for much of the year, becausethe saturation maintains an anaerobic environment that retards decomposition (Everett, 1983).Glaz (1995) suggested that annual durations of saturation required for organic soil accumulationrange from 15 to 94%. Once drained for agriculture, the surface of the organic soils decreases inelevation over time. Processes responsible for the decrease include both primary subsidence andsecondary subsidence (Everett, 1983). As shown in Fig. 1, primary subsidence is a relativelyrapid process that results from a loss of buoyant force following drainage that causes the soil tosink under its own weight. Secondary subsidence is slower, and caused by decomposition(oxidation) of the organic debris as well as shrinkage following drying. Organic soil materialcan also be lost by burning which is a rapid form of oxidation and contributes to secondarysubsidence.

Rates of primary and secondary subsidence are related to the original thickness of thesoil, depth to water table (Stephens, 1954), mineral content (Slusher et al., 1974), temperature,precipitation, and management practices (Shih et al., 1998). Subsidence rates have beendetermined directly using benchmarks and surveying techniques before and after drainage hasoccurred (Stephens, 1954; Shih et al., 1998; Millette, 1976). Mathur et al. (1982) estimatedsubsidence rates of organic deposits from drained and undrained areas using pollen analysis.Unique pollen types or elevated levels of pollen were used as chronological markers to datewhen an organic layer was exposed at the surface. The differences in the depths of thechronological markers were then used to estimate subsidence. Mathur et al. (1982) found thatthe rate of subsidence was 6 cm yr-1 after 5 years and 3.67 cm yr-1 after 15 years.

Dolman and Buol (1967) estimated subsidence in an area in North Carolina that wasdrained for 50 years, but not farmed, using the “one third thickness loss upon drying rule”.Subsidence was estimated by assuming that the thickness of the existing organic soil was onethird of its original thickness in drained areas which were not used for agriculture. Theyextrapolated the estimated original elevation across adjacent agricultural areas and found that thesoils used for agriculture had subsidence rates between 1.8 to 3.54 cm yr-1 . Rates varied withthe thickness of organic layer with the shallower organic layer having the higher subsidencerates.

Several methods been developed to control subsidence through water and landmanagement. Levesque and Mathur (1984) have shown that additions of copper willreduce the rate of subsidence by inhibiting soil enzymes that control the rate of oxidationof organic matter. Covering the surface with mineral material to slow diffusion of

164

oxygen has also been tried with some success (Slusher et al., 1974). Keeping the amountof water in the peat to less than 50% or greater than 80% can slow decomposition(Stephens, 1954). Maintaining a high water table also reduces subsidence (Shih et al.,1998). Brooks and Lowe (1984) predicted that in Florida saturation for 60% of the yearwould slow subsidence of the soils of the Upper St. Johns River.

Wetland restoration projects frequently plug the ditches of drained agriculturalfields and plant trees to recreate original conditions. In areas where organic soils havesubsided, it is possible that ditch plugging will raise the level of groundwater above thesoil surface. Fennema et al. (1994) predicted that the Everglades Agricultural Area(EAA) would be flooded for up to 360 days a year if man-made structures were removed,because the soils are 1.5 m deeper than the original surface. In extreme cases this maykill the newly planted vegetation before it has become established. We hypothesized thatif the amount of subsidence could be predicted before ditches were plugged, then thepotential problem of too great a water table rise following plugging might be avoided.

The amounts of subsidence that have occurred at a site will vary because ofvariations in water table depth, and the effects of past land treatments that were notuniformly applied across the field. Such treatments include drainage, tree removal,crowning of fields, stockpiling and burning of debris, among other things (Lilly, 1981).For these reasons, subsidence predictions should be made at as many locations aspossible across a field. Subsidence processes produce two major changes in an organicsoil: they increase bulk density through settling and also increase the concentrations ofmineral components, such as sand, in horizons. Brewer (1976) showed howmeasurements of bulk density and mineral concentrations could be used to computechanges in volume of soil horizons following weathering and loss of material. Wehypothesized that similar measurements could be made on organic soils to estimateamounts of subsidence using measurements of bulk density and sand percentage. Theobjective of this work was to estimate the amounts of primary and secondary subsidencethat occurred in a drained organic soil using soil property measurements in drained andreference undrained soils.

THEORY AND ASSUMPTIONS

A hypothetical organic soil profile that has undergone both primary andsecondary subsidence has two distinctly different organic soil horizons (Fig. 2). The Oaphorizon is sapric material or muck that has been plowed for agriculture. It has a stronggranular soil structure produced by drying and shrinkage, oxidation, and tillage (Lilly,1981). This layer turns black in color as the organic materials oxidize. The underlyingOa horizon, also sapric material or muck, is below the tillage zone and remains wetterand less oxidized than the Oap horizon. The Oa horizon has a massive soil structure andshows no evidence of shrinkage and drying. It may have a redder color than that of theOap horizon when oxidation is limited, and is massive in structure because no shrinkageoccurred to create cracks. The Cg horizon is mineral soil material and is not of interest inthis study.

165

We assumed that the Oa horizon with a massive structure experienced primarysubsidence but virtually no secondary subsidence. Drying and shrinkage creates soilstructure in the form of well-defined aggregates separated by large cracks (Pons andZonneveld, 1965 and Lee and Manoch, 1974). The Oap horizon was affected by bothprimary and secondary subsidence because it had well developed structure and also ablack color. It was clearly within the depth of plowing and could be easily distinguishedfrom the underlying organic material.

Method for estimating Primary Subsidence

Primary subsidence occurs following drainage when the organic material settlesunder its own weight and compresses (Fig. 3). Volume is reduced by a reduction of porespace. There is minimal loss of organic material. The thickness of the organic soil layerdecreases as it subsides causing bulk density to increase. This change in bulk densitybefore and after drainage can be used to calculate the change in thickness of the organicsoil material as a result of primary subsidence.

Loss of volume through primary subsidence was estimated for a volume oforganic soil that had a unit cross-sectional area and a height that included the entirethickness of the organic soil material. The volume of the original organic soil (Vo) wascomputed as:

Vo = To (1cm2) (1)where To is the thickness of the original organic soil material. The volume of the existingsoil (Vsp) after primary subsidence was computed as:

Vsp = Tsp (1cm2) (2)The mass of the organic soil material before primary subsidence (Mo) is assumed to beequal to the mass of the soil after primary subsidence (Msp). The reduction in volumefollowing primary subsidence occurred only by a decrease in the thickness of the originalsoil volume (To). The cross-sectional area of the soil was assumed to remain unchangedduring primary subsidence.The bulk density of the original soil (Do) is computed as:

Do = Mo / Vo = Mo / To (1cm2) (3)The bulk density for the soil after primary subsidence is computed as: Dsp = Msp / Vsp = Msp / Tsp (1cm2) (4)The ratio of Dsp and Do is related to the ratio of the thickness of the soil layers:

Dsp / Do = [Msp / Tsp (1cm2)] / [Mo / To (1cm2)] = To / Tsp (5)because we assumed that Msp = Mo. The thickness of the original organic soil material

(To) can be estimated as:To = Tsp [Dsp / Do] (6)

The amount of primary subsidence (Sp) that has occurred is the difference between thevertical heights. Sp = To - Tsp = Tsp[Dsp/Do] – Tsp = Tsp[[Dsp/Do] – 1] (7)

In making these calculations we assumed that the drained soil subsides to auniform bulk density in all organic horizons. Values for Dsp and Tsp are obtained fromthe organic soil that had subsided. A value for Do can be obtained by sampling undrained

166

organic soil. The undrained soil is assumed to have a bulk density that is similar to whatthe drained soil had prior to subsidence.

Method for estimating Secondary Subsidence

While organic soils form by the accumulation of organic materials, they usuallycontain small amounts of sand and silt that are deposited by wind and water. The sandand silt consist primarily of quartz, which is a mineral that does not weather or alter overshort time periods of < 1000 years or so (Buol et al., 2003). Oxidation of organicmaterial by either decomposition or fire increases the concentration of mineral material ina soil horizon. Therefore, the concentration of mineral material can be used to estimatethe amount of secondary subsidence that has occurred. Changes in subsidence of anorganic soil following drainage will be discussed using a volume of undrained soil thathas a cross section area of 1 cm2. It is assumed that this area remains constant duringsubsidence, and the volume decrease occurs only through a decrease in horizontalthickness.

We used a method described by Brewer (1976) to estimate the amount secondarysubsidence. Brewer (1976) estimated the soil volume lost when primary mineralsweather. While he applied the technique for the weathering of mineral materials, weadapted it for the decomposition of organic material. Two soil volumes must becompared, the parent material or unoxidized organic soil (Vp), and a volume of oxidizedsoil (Vs). As the parent material weathers (oxidizes) it loses both volume and mass andthe minerals that resist weathering actually increase in concentration as compared to theoriginal material (Fig. 4). This relationship was described by Brewer (1976) as:

VsDsRs / 100 = VpDpRp /100 (8)where Vs = volume of present day soil

Vp= volume of parent material from which soil was derived,Ds = bulk density of present day soil horizon,Rs= percentage by weight of the stable constituent in present day soil horizon,Dp = bulk density of parent material, andRs= percentage by weight of the stable constituent in the parent material.

Changes in the thickness of a horizon can be calculated by assuming that the change involume through weathering occurs only in the vertical dimension and the cross-sectionalarea of the soil remains constant. When a soil mass shrinks it appears that the shrinkageoccurs in all dimensions. However, the shrinkage creates voids in the horizontaldimension, and these are part of the soil volume (Fig. 5). As a result, the soil’s cross-sectional area is assumed to remains constant throughout the weathering process becausesolid material is replaced by voids. Secondary subsidence changes volume by decreasingthe thickness of the soil horizon (Ts):

Vs = Ts (1cm2) (9)The volume of parent material (Vp) can also be expressed as a function of thickness andunit cross-sectional area. (Tp). Equation 8 can then be arranged to calculate Tp.

Tp = Ts * (DsRs / DpRp) (10)

167

Secondary subsidence (Ss) would be equal to the difference in the thickness of the parenthorizon and the soil horizon.

Ss = Tp – Ts (11)

We used the underlying sapric material (Oa horizon), which has undergoneprimary subsidence only, as the parent material for these calculations, and used thepercent sand in the organic soil horizons (Oa and Oap) as the stable constituent thatincreases as a result of oxidation. We assumed that the current Oap horizons formedfrom material similar to that of the underlying Oa horizons, , and we also assumed thatthe water table has been maintained at a level to prevent a significant loss of the materialdue to oxidization from the Oa horizon. We realize that there are many environmental,vegetative, and meteorological situations that factor into the deposition of sand into theorganic material. However, we assume that over time, the sand that is deposited on thesurface of the accreting organic soil would be uniformly distributed through the soil.North Carolina is in the path of hurricanes that have winds that can carry sand over longdistances. Sand deposition that is deposited episodically would be homogenized throughthe profile over time by tree throw and bioturbation. We feel that such an assumption isvalid after examining sand percentages through organic profiles in natural Carolina Bays.Sand percentages varied by less than 5% through the sapric part of the organic profiles(data not shown).

Sample calculations for primary subsidence

Data from one location in the drained Carolina bay (Table 1) will be used toillustrate the calculations. Bulk density (Do) was measured in organic Oa horizons in theundrained natural bays and was found to be 0.25 g cm-3. Bulk density (Dsp) from thesubsided soils used for agriculture was determined to be 0.45 g cm-3 in the Oa4 horizon.The Oa4 horizon was the deepest organic layer in the soil, and also one with lowest bulkdensity. Because we assumed that all horizons had a bulk density of 0.45 g cm-3 prior tosecondary subsidence occurring, then Tsp is equal to the sum or all horizon thicknessesfrom the surface to the bottom of the Oa4 horizon. In this example Tsp is equal to 66 cm.Therefore, using equation 4: To = 66 cm [0.45 g cm-3 / 0.25 g cm-3 ] = 119 cm (12) Sp = 119cm - 66cm = 53 cm (13)This gives us an original thickness of 119 cm and a primary subsidence of 53 cm. Thelength of time for primary subsidence to occur varies, and because we have no data tosuggest how long it took for primary subsidence to occur, we arbitrarily chose 10 years.This would give us a primary subsidence rate of 5.3 cm yr-1 if all the estimated primarysubsidence occurred during the 10-yr. period.

Sample calculations for secondary subsidence

Using the data in Table 1 for calculating secondary subsidence, the organichorizon that had the lowest sand percentage was determined to be the “parent” horizonfrom which the horizons above were formed. For this illustration the Oa4 horizon is theparent material, and the overlying horizons are soil material that have experience varying

168

degrees of secondary subsidence. The change in thickness has to be calculated for eachhorizon above the parent horizon and then summed to determine total secondarysubsidence. Using equation 10 for the Oa3 horizon: Tp = Ts * (DsRs) / (DpRp) (14) Tp = 20 cm * ( 0.45 g cm-3 *9.96/100) / (0.45 g cm-3 * 5.84/100) = 34 cm (15) Ss = 34 cm - 20cm = 14 cm (16)

The Oa3 horizon subsided approximately 14 cm through shrinkage and oxidation.Similar calculations were repeated for the Oa2 and Oa1 horizons using the same parentmaterial. The change in depth for Oa1 and Oa2 horizons were 44 cm and 30 cm, and thetotal amount of secondary subsidence across the three horizons was 88 cm. This samplingarea had been drained for 30 years and accordingly secondary subsidence lowered thesurface at a rate of 2.9 cm yr-1.

This method did not work for locations in which the lowest amount of sand in theorganic soil was in the surface horizon because it produced negative values ofsubsidence. Such values implied that there was an accumulation of material, probablydue to spoil added to the soil surface through ditch cleaning. In cases where the sandcontent in soil horizons (Rs) were more than three times the sand content in the parenthorizons (Rp), it was assumed that the soil horizon was fill material (from ditchexcavation) and therefore that soil horizon was not used in the calculations. In additionto using total sand for the secondary subsidence calculations, we tried using the coarseand fine fractions of sand. It was found that using total sand gave an estimate that was amedian of the two sand fractions and so total sand was used for all calculations.

MATERIALS AND METHODS

Juniper bay is a 256 ha Carolina bay southeast of Lumberton, North Carolina(34o30’30”N 79o01’30”E) that was drained and placed into agricultural production over a30-yr period. The Robeson County Soil Survey showed that approximately 60% of theJuniper Bay consisted of organic soils that were classified as members of the Ponzerseries (loamy, mixed, dysic, thermic Terric Haplosaprists) (USDA-SCS, 1978). Organicsoil layers ranged from 40 to 130 cm. Undrained organic horizons have a massive soilstructure, but develop subangluar blocky or granular structure following drainage(official series description). Plant communities believed to have been present in JuniperBay included a Nonriverine swamp forest, High pocosin, and Peatland Atlantic WhiteCedar forest communities as described by Schafale and Weakly, 1990.

Aerial photographs and interviews with previous landowners were used toestablish the drainage history of Juniper Bay. Approximately one third of the Bay wasdrained in 1971, another third drained in 1981, and the last third drained in 1986.Drainage ditches were dug to depths of 1 m in most of the Bay, but these fed into largerditches that extended to depths of approximately 4 m. Field “cuts” were the land areassurrounded by ditches on all sides, which were used for agriculture. At selected fieldcuts, paired sampling points, with one at the crest and one near a ditch, were placed atlocations determined by an equilateral triangle grid randomly placed across the site (Fig.6). Pits were dug to a depth of approximately 1 m using a backhoe. In each pit, the soil

169

profile was described to determine depth, color, and structure of major soil horizons.Uhland cores (75 cm3) were taken from each horizon to determine bulk density. Bulksamples were taken from each horizon described in the profile. Total organic carbon wasdetermined through dry combustion with a Perkin-Elmer PE2400 CHN ElementalAnalyzer. Particle size was determined with the pipette method on a 10 g sample.Samples were prepared for particle size analysis by first oxidizing the organic matter withconcentrated (30%) hydrogen peroxide. The amount of mineral material in the initial 10g sample was determined with an additional 10 g sample that was placed in a mufflefurnace at 400oC for 24 hours to remove organic matter, leaving the mineral material.Percent sand from the total sample was calculated by multiplying the percent sand fromthe particle size analysis by the mass of the mineral material and then dividing by themass of the original sample.

Three natural undrained Carolina bays, Tatum Millpond Bay (34o43’00”N78o33’00”E), Charlie Long Millpond Bay (34o46’00”N 78o33’30”E), and Causeway Bay(34o39’45”N 78o25’45”E), were selected for comparison in Bladen County, NorthCarolina. These natural bays have organic soils that are classified as a Pamlico series(sandy or sandy-skeletal, siliceous, dysic, thermic Terric Haplosaprists). These soils haveorganic material 16 to 51 inches thick and are underlain by sand. Plant communitiesfound in these bays were Pond Pine Woodland, Non- riverine Swamp Forest, Bay Forest,and High Pocosin, as described by Schafale and Weakly (1990).

Trails were cut through dense vegetation into the natural bays to reach samplinglocations. Soil profiles at four locations in the organic soils of each bay were describedusing a McCauley peat sampler. Bulk samples were taken from each horizon for carbonand particle size analysis. Bulk density was determined by taking a 10 cm undisturbedsample with the McCauley peat sampler.

The initial study on Juniper Bay was not designed for determining subsidence of organicsoils, and because of this we were unable to perform a statistical analysis of the results.Data reported consist of calculated estimates and averages.

RESULTS AND DISCUSSION

The organic soils from Juniper Bay and the natural bays were classified as TerricHaplosaprists. Typical profile descriptions are shown in Table 2, and selected physicaland chemical properties are shown in Table 3 for one soil the drained site and another ina reference site. The undrained organic soils had a surface Oi and Oe horizons thatcombined were 50 cm thick, while these horizons were absent in the drained CarolinaBay. The Oi horizons were composed of fibric material that consisted of dead leaves anddead roots from the surrounding trees. Oe horizons were similar but had undergone moredecomposition. Removal of the natural vegetation and use of the field for agricultureprevents Oi and Oe horizons from developing.

The Oa horizons in the undrained bay had massive structure. In contrast thedrained Carolina Bay had an Oap horizon at the surface with strong granular structure asa result of tillage and dessication. In some sampling locations, very coarse prismatic to

170

subangular blocky structure had developed in the Oap horizons. The Oa horizons in thedrained bay were very dark brown (10YR 2/2) sapric (muck) compared to the black(10YR 2/1) sapric material of the undrained bays. The organic horizons were thicker (upto 170cm thick) in the undrained bay, compared to 52 cm in the drained bay.

Bulk density and sand percentage were two to three times higher in the organichorizons of the drained bay (Table 3). The difference in bulk density and changes instructure between the drained and undrained bays were the result of primary subsidence.The amount of sand in the organic soils of the drained bay was higher in at the surfaceand decreased with depth, while the amount of sand in the Oa horizons of the undrainedbay was relatively constant. This trend in sand percentage change is what we expected tosee in areas where secondary subsidence had occurred. Not all organic soil locations thatwere sampled could be used for analysis. Plots were deleted when the lowest sandpercentage was in the surface horizon, or if there was no Oa horizon that could be used asparent material. The locations that were chosen are shown in Table 4.

Estimates for total subsidence varied among locations and with time sincedrainage (Table 4). Averages of the total subsidence values across the three time periodsshow that the organic soils subsided approximately 80 cm. Subsidence in the fieldsdrained for 15 years is slightly less, 75 cm, than for fields drained for 20 or 30 years, 77cm 86 cm respectively. This trend of total subsidence is expected indicating that thelonger a site is drained the greater the amount of subsidence occurs. These values fortotal subsidence pertain to the Oa and Oap horizons which were sampled. We did notmeasure properties of the Oi and Oe horizons because they had no counterparts in thedrained organic soil.

Estimated rates of primary and secondary subsidence varied across locations(Table 4). The average rate of primary subsidence is 4 cm yr10

-1. Primary subsidencerates were lowest in the areas drained for 15 years, 3 cm yr10

-1, and were highest in theareas drained for 20 years 4 cm yr10

-1. This is probably due the water table in the areadrained 15 years ago is closer to the surface relative to the other areas, maintainingbuoyancy over more of the profile. Shih et al. (1998), showed that the rate of subsidencein the EAA have decreased from 2.5-3.0 cm yr-1 during the years 1913 to 1978, to 1.45cm yr-1 during the years 1978 to 1997 as a result of better water management whichraised the water table to slow decomposition. The average rate of secondary subsidencewas approximately 2 cm yr-1 (Table 4). Secondary subsidence rates were highest after 15years of drainage, 3 cm yr-1, and similar after 20 and 30 years of drainage atapproximately 2 cm yr-1.

These estimates of subsidence and subsidence rates are consistent with otherreported subsidence rates in previous studies. Stephens (1956) reported a subsidence rateof 4.3 cm yr-1 over a 50 year period in Florida, Ireyresr (1963) reported a rate of 1.5 cmyr-1 over 100 year period in England, and Jongedyk et al., (1950) reported a rate of 1.5cm over 6 years in Indiana. Our calculated values are also similar to those found by Tant(1979) in North Carolina. A yearly loss of 1.2 cm was found on a Belhaven muck and0.38 cm on a Pungo muck. Average annual subsidence in Quebec was 2.1 ± 0.4 cm yr-1

171

over 38 years (Millette, 1976), and subsidence in New Orleans ranges from 1 to 5 cm yr-1

(Slusher et al., 1974).

The proportion that primary subsidence comprised of the total subsidence variedwidely (Table 4), but overall the average proportion of primary and secondary wereapproximately equal. Largest variations occurred in the field drained for 30 years whereprimary subsidence accounted for between 13 and 85% of total subsidence. This resultwas unexpected because it was thought that secondary subsidence was a slower process.Previous landowners indicated, however, that fire was used in the clearing process toremove tree debris during the clearing operation, and charcoal was found in soil profilesat Juniper Bay (Table 1). Fires also occurred naturally through lightning strikes. If a siteburned shortly after drainage ditches were installed, it is possible that loss throughsecondary subsidence would exceed that of primary subsidence. Loss of the organicmaterial through burning would concentrate sand, and also lessen the weight of thematerial compressing the parent material. This would keep the parent material’s bulkdensity low. Several studies have shown that subsidence tends to decrease withincreasing distance from a ditch (Burke, 1963; Brandof, 1992), however our estimatesneither verified or contradicted their findings when comparisons were made betweenditch and crest locations.

Overall, the results did not show greater subsidence near ditches as expected.This could be due to the ditches being shallow in most parts of the site which kept soilswetter and slowed the subsidence processes. The land clearing process is also complexand involves the use of fire, ditch construction and maintenance, and crowning of fields.Considering the variety of operations that are used to prepare and maintain the land inagriculture it is not surprising that we are unable to see a clear effect of the ditches onsubsidence rates.

SUMMARY AND IMPLICATIONS

The principal finding from this study is that if the drainage ditches were filled into restore wetland hydrology in Juniper Bay, we estimate that the water table will rise toapproximately 80 cm above the existing organic soil surface. This is determined from theaverage value of subsidence that was estimated to have occurred across the site (Table 4).There may be some variation in water depth across the site, because soils drained for 15years have experienced less subsidence. However, as shown in Table 4 the amount ofestimated subsidence within the drainage groups was large and so little consistentdifference would be expected across the site. In support of our estimate for the amount ofsubsidence that has occurred we compared the existing elevation of the organic soils inJuniper Bay to the elevation of the mineral surface lying around the Bay’s perimeter.The mineral soils around the edge of the Carolina bay are higher in elevation than theorganic soils by approximately 60 cm. Previous studies have shown that the surface oforganic soils on broad flats can actually be higher than the surfaces of adjacent mineralsoils (Daniels et al., 1999). We feel that for restoration planning it is reasonable toassume that 80 cm of subsidence has occurred is justified. This value does not include

172

the thickness of Oi or Oe horizons which may have occurred over the Oa horizons whenthe natural vegetation covered the site.

We realize that our method for estimating the amount of subsidence that hasoccurred is based on several assumptions that will not hold true for other sites. Potentialsources of error for the results from this study include the following. Some samplelocations near ditches had the organic materials contaminated by ditch cleaningoperations, leading to increased sand at the surface. This would cause increasedestimates of secondary subsidence. The only way to avoid this problem is throughcareful sampling and eliminating sites with extraordinarily high sand percentages. Ourassumptions that original sand percentage and bulk density between the parent materialand overlying layers were similar could also be in error. This problem can be minimizedwhen the parent material properties are selected separately for each sampling site. If onesampling location does not have sufficient undisturbed material in a lower horizon, thenthat site should not be used to estimate subsidence. The impact of fire on increasing theorganic soil oxidation rate is unclear, but has been assumed to result in an increase insand percentage in the burned layers. We feel that these potential sources of error havehad only a small impact on our results because the data reported in Table 4 are in linewith published findings from elsewhere. The best way to estimate subsidence amounts iswith long-term, permanent bench marks whose initial elevation is known. When theseare not available, then our method seems to offer a workable solution to approximatingsubsidence at many sites.

173

REFERENCES

Brandof, K.L. 1992. Impact of ditching and road construction on Red Lake Peatland. InThe Patterned Peatlands of Minnesota. H.E. Wright, Jr., B.A. Coffin, and N.E.Aasenged. University of Minnesota Press, Minneapolis.

Brewer, R. 1976. Fabric and Mineral Analysis of Soils. Robert E. Krieger PublishingCompany; Huntington NY. 63-87.

Brooks, J.E., and E.F. Lowe. 1984. U.S. EPA clean lakes program, phase I. Diagnosticfeasibility study of the Upper St. Johns River Chain of Lakes. Vol II-feasibilitystudy. Technical Publications SJ 84-15, St. Johns River Water ManagementDistrict, Palatka, Florida.

Buol, S.W., R.J. Southard, R.C. Graham, P.A. McDaniel. 2003. Soil Genesis andClassification 5th ed. Iowa State Press. Ames, Iowa.

Burke, W. 1963. Drainage of blanket peat at Glenamory. In Proceedings of the SecondInternational Peat Congress, Leningrad, USSR. p 809-817.

Collins, M.E. and R.J. Kuehl, 2001. Organic matter accumulation and organic soils. InWetland Soils: Genesis, Hydrology, Landscapes, and Classification. J.L.Richardson and M.J. Vepraskas ed. Lewis Publishers, New York. 137-162.

Daniels, R. B., S. W. Buol, H. J. Kleiss, and C. A. Ditzler. 1999. Soil systems in NorthCarolina.Technical Report 314, Department of Soil Science, North CarolinaState University, Raleigh.

Dolman, J.D., and S.W. Buol. 1967. A Study of Organic Soils (Histosols) In theTidewater Region of North Carolina. North Carolina Ag. Ex. St. Tech. Bull. No.181.

Everett, K.R. 1983. Histosols. In Pedogensis and Soil Taxonomy II. The Soil Orders.L.P. Wilding, N.E. Smeck an dG.F. Hall eds. Elsevier, Amsterdam. pp. 1-53.

Fennema, R.J., C.J. Neidrauer, R.A. Johnson, W.A. Perkins, and T.K. MacVicar. 1994.A computer model to simulate natural Everglades hydrology. In Davis, S.M., andJ.C. Ogden (eds.), Everglades: The Ecosystem and its Restoration. St. LuciePress, Delray Beach, Florida.

Glaz, B. 1995. Research seeking agricultural and ecological benefits in the Everglades.J. Soil and Water Conservation. 609-613.

174

Ireyresr, 1963. Vegetation and Soils, A World Picture. Aldine, Chicago. pp 324.Jongedyk, H.A., Hickok, R.B., Mayer, I.D. and N.K. Ellis, 1950. Subsidence ofmuck soils in northern Indiana. Purdue Univ. Agric., Exp. Sta., Spec. Circ., 366:1-10.

Lee, G.B. and B. Manoch. 1974. Macromorphology and Micromorphology of aWisconsin Saprist. In Histosols Their Characteristics, Classification and Use.M. Stelly and R.C. Dinauer ed. SSSA pub 6. Soil Sci. Soc. Am. Madison,Wisconsin.

Levesque, M. P. and S.P. Mathur. 1984. The effect of using copper for mitigationHistosol subsidence on : 3. The yield and nutrition of minicarrots, carrots, andonions grown in Histosols, mineral sublayers, and their mixtures. Soil Sci.138:127-137.

Lilly, J.P., 1981. The Blackland Soils of North Carolina: Their Characteristics andManagement for Agriculture. North Carolina Ag. Res. Tech Bul. No. 270. pp. 27-33.

Mathur, S.P., M.P. Levesque, and P.J.H. Richard. 1982. The establishment of synchronybetween subsurface layers and estimation of overall subsidence of cultivatedorganic soils by a palynological method. Can. J. Soil Sci. 62:427-431.

Millette, J.A. 1976. Subsidence of an organic soil in southwestern Québec. Can. J. Soil.Sci. 56:499-500.

Pons, L.J. and I.S. Zonneveld. 1965. Soil Ripening and Soil Classification Initial SoilFormation in Alluvial Deposits and a Classification of the Resulting Soils.International Institute for Land Reclamation and Improvement pub 13.H.Veenman & Zonen N.V. Wageningen, Netherlands.

Schafale, M.P. and A.S. Weakley. 1990. Classification of the Natural Communities ofNorth Carolina: Third Approximation. North Carolina Natural Heritage Program,Division of Parks and Recreation, Department of Environment, Health andNatural Resources. Raleigh, NC. p. 205-217.

Shih, S.F., B. Glaz, and R.E. Barnes, Jr. 1998. Subsidence of organic soils in theEverglades Agricultural Area during the past 19 years. Soil and Crop Sci. Soc.Florida.57:20-29.

Slusher, D.F., W.L. Cockerham, and S.D. Matthews. 1974. Mapping and interpretationof Histosols and Hydraquents for urban development. In Histosols; TheirCharacteristics, Classification, and Use. SSSA Special Pub 6. A.R. Aandalh, S.W.Buol, D.E. Hill, and H.H. Bailey eds. Soil Sci. Soc. Am. Inc., Madison, WI. 95-109.

175

Stephens, J.C. 1956. Subsidence of organic soils in the Florida Everglades. Soil Sci.Soc. Am. Proc. 20:77-80.

Tant, P. 1979. Subsidence of organic soils. Soil Sci. Soc. N.C. Proc. 22:75-80.USDA-SCS. 1978. Soil Survey of Robeson County, North Carolina.

176

Table 1. Values for sample calculations. Bulk density from Oa4 is used for estimatingprimary subsidence because it is the lowest. Horizon Oa4 is used as the parent materialin estimating secondary subsidence because of the lowest % sand. All other horizonshave a concentration of sand that could be attributed to secondary subsidence.

HorizonHorizon

ThicknessBulk

Density Sand Change in horizon thickness

(cm) (g cm-1) (%total sample) (cm)

Oa1 10 0.75 18.7 43.6Oa2 10 0.75 13.9 12.7Oa3 20 0.45 10.0 14.1Oa4 26 0.45 5.8 0 (parent material)

177

Table 2. Typical profile from an organic soil at an undrained natural Carolina bay andfrom a drained Carolina bay.

Horizon Depth Descriptioncm

Undrained

Oi 0-20 Black (10YR 2/1) fibric to hemic muck. Massive structure.Gradual boundary. Root and debris mat.

Oe 20-50 Very dark brown (7.5 YR 2.5/2) hemic muck. Massivestructure. Gradual boundary. Organic bodies 0.5-1 cm. Many

roots and debris.

Oa1 50-82 Very dark brown (10YR 2/2) sapric muck. Massive structure.Clear boundary. Common large pieces of wood debris.

Oa2 82-107 Black (10YR 2/1) sapric muck. Massive structure. Gradualboundary. Few large pieces of wood debris.

Oa3 107-145 Black (10YR 2/1) sapric muck. Massive structure. Gradualboundary

OC 145-170 Black (10YR 2/1) sapric muck with <2% sand grains. Massivestructure. Abrupt boundary.

C 170-190 Very dark brown (10 YR 2/2) sand. Single grain structure.

Drained

Oap 0-11 Black (N 2.5/0) sapric muck. Strong medium (2mm) granularstructure. Abrupt boundary.

Oa1 11-31 Very dark brown (10YR 2/2) sapric muck with 10% woodfragments and 10% charcoal. Massive structure. Diffuse

boundary.

Oa2 31-52 Very dark brown (10YR 2/2) sapric muck with 10% woodfragments and 10% charcoal. Massive structure. Abruptboundary.

Bw 52-61 Dark reddish brown (5YR 3/3) sandy loam. Moderate coarse(5cm) prismatic structure. Abrupt boundary.

2BC 61-78 Yellowish brown (10YR 5/4) loamy sand with 40% white (2.5Y8/1) sand in areas 1mm in diameter. There was 10% woodfragments. Massive structure. Clear boundary.

178

2C1 78-100 Light yellowish brown (10YR 6/4) loamy sand with 5% woodfragments. Massive structure. Clear boundary.

2C2 100+ Dark gray (10YR 4/1) loamy sand. Single grain structure.reaction to alpha, alpha. Dense.

179

Table 3. Particle size, organic carbon, and bulk density of typical profiles from Juniperbay and undrained natural bays.

Horizon depth C BD sand silt clay

(cm) (%) g cm-3 ----------(%)----------

DrainedOap 0-11 26.25 0.76 34.2 14.3 10.3

Oa1 12-31 52.52 0.47 16.7 22.9 7.3

Oa2 32-52 27.91 0.47 41.1 18.1 7.8

Bw 53-61 5.46 0.93 74.0 8.5 9.6

BC 62-78 0.48 1.55 93.1 1.8 4.3

C1 79-100 0.26 1.55 91.3 2.5 5.9

C2 101-110 0.32 --- 91.9 1.7 5.9

UndrainedOi 0-20 41.25 --- --- --- ---

Oe 21-50 37.37 --- --- --- ---

Oa1 51-82 42.28 0.18 6.8 11.5 9.0

Oa2 83-107 34.74 0.21 14.0 17.3 10.0

Oa3 108-145 34.56 0.19 14.2 29.9 10.3

OC 146-170 16.35 0.49 46.1 20.4 6.8

C 190 --- 0.87 78.7 12.1 1.7

Table 4. Estimated secondary subsidence from selected locations. The C or D in thesample location indicates if the sampling pit was at the crest (C) or near the ditch (D).Rate for secondary subsidence was calculated by dividing the amount of subsidence by

180

the years of drainage (x). Primary subsidence rate was estimated by dividing the amountof subsidence by 10 years. * Did not include surface horizons that had greater that threetimes Rp in the calculation of secondary subsidence.

Primarysubsidence

Secondarysubsidence

Total subsidencePlotLocation

Amount Rate Amount Rate Absolute Primary Secondary

(cm) (cm/yr10) (cm) (cm/yrx) (cm) ------ (%) -----

15 Years After Drainage

66D* 13.9 1.4 42.3 2.82 56.2 25 7568D 26.8 2.7 55.0 3.67 81.8 33 678D 60.2 6.0 26.8 1.79 87.0 69 31Av. 33.6 3.4 41.4 2.76 75.0 42 58

20 Years After Drainage

10C* 24.4 2.4 50.0 2.50 74.4 32 6816C 45.8 4.5 25.3 1.26 71.1 65 3516D 54.3 5.4 45.2 2.26 99.5 54 466C* 72.0 7.2 23.6 1.18 95.6 75 2511C 20.4 2.0 21.8 1.09 42.2 48 52Av. 43.4 4.3 33.2 1.66 76.6 57 43

30 Years After Drainage

2D 33.6 3.3 6.2 0.21 39.8 84 164D 61.4 6.1 8.3 0.28 69.7 87 135C 52.8 5.2 85.9 2.85 138.7 43 575D 14.4 1.4 82.9 2.76 97.3 15 85Av. 40.6 4.0 45.8 1.52 86.4 47 53

181

Table 5. Mean and range of bulk density and total sand in the surface two horizons oforganic soils in Juniper Bay (JB) and the Reference Bays (RB). Notice the inverserelationships in bulk density and total sand of the surface two horizons between JuniperBay and the Reference Bays. NOTE: The mean and range values include horizons thatwere excluded in making subsidence calculations.

Bulk Density Total sand----- (g cm-3) ----- ----- (%) -----mean range mean range

JB 1 0.77 0.53-1.23 49.0 24.1-88.0JB 2 0.61 0.32-1.23 41.9 13.2-87.3RB 1 0.25 0.10-0.74 36.7 11.3-64.0

RB 2 0.52 0.16-0.87 54.8 53.3-83.1

182

Fig. 1. Hypothetical illustration of the effects of subsidence on the decrease in elevation of anorganic soil following drainage. Primary subsidence occurs quickly due to loss of waterthrough drainage. Secondary subsidence is a slower process.

Time

Surf

ace

Ele

vati

on (m

)

0

2

1

Drainage Begins

Primary subsidence: lossof buoyant force

Secondary subsidence: lossby oxidation and shrinkage

183

Oap

Oa

Cg

Horizon has stronggranular structure andhigh sand concentration

Horizon has massive structure,and low sand concentration.Oap formed from this material.

Mineral soil material notrelated to overlying horizons

Fig. 2. Example of a soil profile consisting of organic soil material overmineral material. The Oap horizon is tilled and undergoes the mostoxidation. The Oa horizon showed little influence of oxidation and was usedas the parent material for the Oap horizon. The mineral material was notconsidered in the calculations.

184

Sand grain

Volume of an organic layer before drainage

Volume after lossof buoyant forces

Vsp

Vo

Fig. 3. Illustration of how primary subsidence reduces volumewhile the mass remains the same in an organic soil horizonfollowing loss of buoyant forces. Vo is the volume of soilhaving a unit cross-section before primary subsidence occurs.Vsp is the volume of soil after primary subsidence, which alsohas a unit cross-sectional area. It is assumed that subsidencehas only altered soil volume in the vertical direction.

185

Sand grain

Volume of an organic layer after drainage but before secondary subsidence

Volume after lossdue to oxidization

Vp ≅ Vsp

Vsp Vs

Fig. 4. Illustration of how secondary subsidence due tooxidization reduces volume while also increasing the mass inthe surface layers by concentrating the sand. Vsp is the volumeof soil after primary subsidence but before secondarysubsidence. Vp is the volume of parent material that has notundergone secondary subsidence and is approximately equalto Vsp. Vs is the volume of soil that has undergone secondarysubsidence. It is assumed that all volumes have a unit cross-sectional area and changes in volume occur in the verticaldirection.

186

Fig 5. When a soil mass shrinks it appears that the shrinkageoccurs in all dimensions. However in the horizontal dimensionthe shrinkage creates voids, and these are part of the soilsvolume. As a result, the soil’s cross-sectional area is assumedto remain constant throughout the weathering process, withchanges occurring in the vertical dimension.

Width (Wpm)

Thickness(Tpm)

Width (Ws)

Thickness(Ts)

Cracks formed byshrinkage

Wpm=WsTpm>Ts

Undrained Organic Soil(Parent Material)

Drained Organic Soil

187

Fig. 6. Schematic diagram showing the drained Carolina Bay (Juniper Bay) withits drainage ditches. Sampling locations within the organic soils are shown bydots. At each dot, one pit was located at the center of the field cut, equidistantbetween ditches while another pit was placed near (within 2 m) of a ditch.

200 0 200 400 600 800 1000 Meters

N

##

###

##

#

##

##

#

##

##

#

##

###

####

####

#

####

#

Ditches

188

Chapter 8

NUTRIENT ANALYSIS OF DRAINAGE WATERS FROM JUNIPER BAY

G.S. Kreiser, M.J. Vepraskas, and R.L. Huffman

INTRODUCTION

Many points through the nutrient cycle are strongly influenced by the hydrologiccycle (Bormann and Likens, 1967). Inputs and outputs of nutrients are directly related tothe amounts of water that move in and out of an ecosystem. The biological uptake ofnutrients by plants and the release of nutrients by decomposition are related to the amountand pattern of water availability. The rate and nature of weathering and soil formationare influenced by the hydrologic regime, since water is essential to major chemicalprocesses (Bormann and Likens, 1967).

Nutrient cycling in wetlands can be defined in the context of water budgets. Thesource, velocity, quantity, and distribution of water directly control the spatialheterogeneity of the nutrients (Carter, 1986). However, relatively few studies have beenperformed on nutrient exports and imports based on water budget studies (Carter, 1986).

Water quality in unaltered wetlands reflects the quality of the water into the wetlandand the interaction of this water with the soils and vegetation (Carter et. al., 1978).Wetlands affect water quality through element cycling, sediment deposition, and ion andmolecule adsorption (Carter, 1986). Wetlands can act as a filter, where the quality of thewater leaving the wetland may differ significantly from that of the water entering thewetland.

Carolina Bays are the only abundant lentic systems of natural origins in the coastalplains of North Carolina (Newman and Schalles, 1990). Lentic is defined as beingassociated with still water systems. These systems occur in basins and lack a definedchannel or floodplain (Alberta Natural Heritage Information Centre, 2002). Even thoughCarolina Bays are very numerous, Schalles and Shure (1989) state, “Carolina Bays arepoorly studied with respect to hydrology, community structure and succession, trophicdynamics, and mineral cycling.” Carolina Bays typically have water chemistries that aresoft and acidic. Biological production is low to moderate (Sharitz and Gibbons, 1982;Newman and Schalles, 1990).

The United States Environmental Protection Agency (USEPA), in its 1996 NationalWater Quality Inventory, claimed that 40 percent of streams or rivers surveyed wereimpaired because of the nutrients N and P (USEPA 1998). Concerns about excessnutrients include adverse effects on humans and domestic animals, aesthetic impairments,negative impacts on aquatic life, and excessive nutrient input into downstream systems(Dodds and Welch, 2000). Eutrophication, which is the addition of excessive nutrients,can cause proliferation of algal masses that can cause degradation of water quality(Dodds and Welch, 2000).

189

Sources for pollution come from either point or non-point sources. Non-pointsources originate from agriculture activities (fertilizer and manure application to land),and residential activities (on-site waste disposal) (Osmond et. al., 1995). Point sourcesinclude industries that use nitrates in manufacturing and sewage treatment plants. In thesoutheastern United States, the main source of nitrogen pollution comes from agriculturalfertilizers and animal manure (Osmond et. al., 1995). The limit suggested for nitrate indrinking water is 10 mg/L (Osmond et. al., 1995). Adverse health effects of nitrate abovethis limit include methemoglobinemia (blue baby syndrome).

Phosphorus in freshwater systems exists either in a particulate or dissolved phase.Particulate matter includes living and dead aquatic organisms, precipitates ofphosphorous, and amorphous phosphorus. The dissolved phase includes inorganicphosphorus, which is generally in the soluble orthophosphate form, and organicphosphorus (Osmond et. al., 1995).

The EPA water quality criteria for phosphates states that levels should not exceed0.05 mg/L if streams discharge into lakes, 0.025 mg/L within in a lake, and 0.1 mg/L instreams or flowing water not discharging into lakes in order to control algal growth(Osmond et. al., 1995). Phosphates do not have notable adverse health effects (USEPA1986). The primary anthropogenic nonpoint sources of phosphorus include runoff fromland areas being mined for phosphate, agricultural areas, and urban/residential areas.Point sources of phosphorus include sewage treatment plants, and industrial wasteproducts (Osmond et. al., 1995).

Many bodies of freshwater are experiencing influxes of both phosphorus andnitrogen from point and nonpoint sources. The increased concentrations of availablephosphorus allows for more assimilation of nitrogen before the phosphorus is depleted.Although levels of orthophosphate between 0.08 to 0.10 mg/L may trigger periodic algaeblooms, long-term eutrophication will be prevented if total phosphorus andorthophosphate levels are below 0.5 mg/L and 0.05 mg/L, respectively (Osmond et. al.,1995).

The environmental effects on freshwater systems of the addition of N and P are thatthe growth of macrophytes and phytoplankton is stimulated. Generally, phosphorus (asorthophosphate) is the limiting nutrient in freshwater aquatic systems. If all phosphorusis used, plant growth will cease, no matter how much nitrogen is available (Osmond et.al., 1995). The natural levels of orthophosphate usually range from 0.005 to 0.05 mg/L.

The objectives of this study were to examine a drained Carolina Bay that was inagricultural production in order to determine: 1) the possible inputs and outputs ofnutrients into the bay; 2) the overall nutrient status of the bay; and 3) if the nutrientsleaving the bay in surface outflow could be a possible source for water qualitydegradation.

190

MATERIALS AND METHODS

The North Carolina Department of Transportation (NCDOT) is in the process ofrestoring a drained Carolina Bay in Robeson County. The project will providecompensatory wetland mitigation in the Lumber River Basin of southeastern NorthCarolina, which will offset wetland impacts from road construction projects in the riverbasin (Hauser, 2001). The goal of the restoration project is to restore the functions andvalues of a Carolina Bay. As part of this restoration project, North Carolina StateUniversity is investigating the hydrologic, soil and vegetative changes that occur inJuniper Bay as a result of this restoration project.

The site, known as Juniper Bay, is composed of 300 hectares of an extensivelydrained Carolina Bay that was limed and fertilized to enhance agricultural production.The nutrient enrichment of Juniper Bay raises the concern that the increased nutrientsmight lead to high levels of nutrients exiting in the surface outflow from the bay andcause water quality degradation.

Nutrient analysis was performed on bulk precipitation, surface outflow andgroundwater. Water samples were sent to the Soil Science Analytical ServiceLaboratory where they were tested for NH4

+-N, NO3--N, PO4

3-, TOC, Ca2+, K+, Mg2+, andNa+. Phosphate, NH4

+-N, NO3--N, and TKN were measured using Lachat Quickechem

8000 slow injection auto analyzer. TOC was measured using a Total Organic Analyzer.Ca, K, Mg, and Na were measured using ICP (Inductively Coupled Plasma - Massspectrometry).

Bulk Precipitation

Bulk precipitation was collected at three locations, next to the rain gauges inJuniper Bay (Fig. 1). Bulk precipitation was collected weekly and combined monthly foreach location. The samples were collected following the system of Likens et al. (1967)where there is a funnel to collect precipitation, tubing and reservoir (Fig. 2). Thereservoir container is isolated from the atmosphere by the vapor barrier placed in thetubing. The vapor barriers were used in order to eliminate concentration of solutes byevaporation (Johnson and Swank, 1973). Samples that were contaminated with leaves,bird feces, etc. were discarded. A dilute hydrochloric acid solution was added to thereservoir as a preservative (Radtke, 1999).

Surface Outflow

Surface outflow samples of 100 mL were taken with an ISCO 3700 sampler, whichtook samples about every 48 minutes. Each bottle could hold ten samples with about 3bottles a day being collected. Hydrochloric acid was added to the bottles as apreservative. Bottles were collected weekly and stored at 4oC until filtering and testing.Weekly pH readings were made of the surface water using an Accumet AP62 pH/mv/Ionmeter.

191

At the laboratory, the bottles were combined based on outflow data. Bottles thatrepresented low flow events, rising limb, or peak of events were sampled together.Outflow measurements were plotted against time and the bottles’ sample time was plottedagainst time. By using the graph, individual bottles were then placed together for testing(Figure 3). The point of this technique was to sample storm events in order to testwhether there is a difference in nutrient levels in the surface outflow during these events.Samples were filtered through a 0.45µm filter to remove particulate material that wouldinterfere with the analysis.

Ground Water

Groundwater samples were taken monthly from piezometers located near cores 1, 3,17 and 25 (Fig. 4). These locations were selected because it is thought that theselocations might be the general area of groundwater inflow or outflow at the bay.

Samples of groundwater were retrieved using a hand bailer. Groundwater pH wastaken in the field by using an Accumet AP62 pH/mv/Ion meter. Dilute hydrochloric acidwas added as a preservative and samples were stored in cold room until analyzed.Samples were filtered using a 0.45 µm filter, in order to remove any sediment.

RESULTS AND DISCUSSION

Surface Outflow

Over a period of one year the surface outflow nutrient levels did not significantlydiffer. The only instance of differences in nutrient concentrations in surface outflowoccurred during large rain events. In particular, there was a dramatic increase of nutrientswhen there were more than 5 cm of rainfall in a day. To illustrate this point, nutrientlevels from several periods of no rainfall where averaged together. Table 1 lists theaverage nutrients for these periods. As shown by Table 1 the nutrient levels in thesurface outflow are in general quite low.

To compare the effect of rainfall on the nutrient concentrations, several stormevents of less than 5 centimeters a day were averaged together. These storm eventsrepresent rainfall as little as 0.13 cm to as much as 4.5 cm in a day. Table 2 shows thatwhen there were rain events of less than 5 cm a day, the nutrient concentrations weresimilar to the periods of no rainfall. This indicates that rain events of less than 5 cm donot have a significant impact on the nutrient concentrations in surface outflow.

There were two rain events that had amounts greater than 5 cm of rainfall in a day.During these periods, the nutrient concentrations increased dramatically in the surfaceoutflow (Table 3). For most nutrients, the concentrations increased by at least two timesthe non-rainfall values. Some of the biggest increases in nutrient concentrations includeTOC and calcium -- an increase of four times the concentrations in base flow conditions.

The pattern of nutrient increase that occurs is that a couple of days after the largerain event the nutrient concentrations peak. The nutrient concentrations in the surface

192

outflow stay elevated for several days after the rain event. Figures 5 and 6 show TOCand calcium levels in the surface water before and after a rain event greater than 5 cm in aday. As seen in the figures it takes about a week for the nutrient concentrations in thewater to drop back down to base flow conditions.

A possible explanation of this phenomenon is that during heavy rainfall events, thewater entering the soil is greater than the infiltration rate of the soil, and runoff occurs.The extra water carries the higher nutrient levels in the topsoil and enters the ditchnetwork raising the nutrient levels of the surface water. The slow response of the nutrientlevels to return to base flow conditions is that Juniper Bay is quite large and it takes awhile for the surface runoff to reach the outflow point. Figure 7 shows a possiblescenario with the different color areas representing a possible time frame for the drainageof the bay. The area closest to the outflow point would drain the fastest and areas furtheraway (in the red) from the outflow point would take up to a week to drain. This delay indrainage would result in the long time frame where nutrient levels would remain elevatedin the surface outflow.

A comparison of nutrient levels by season was also made (Table 4). There was nosignificant difference in the nutrient amounts based on season. A simple test of ANOVAwas performed on all the nutrients by season and the test reveals that there is nodifference in the means. More significantly, it indicates that there are no differences innutrient levels throughout the year, even with the large rainfall events that occurred in thesummer and in the fall.

pH

The pH was measured weekly at the surface outflow. Throughout the year the pHof the surface water ranged from 3.7 to 6.4 (Fig. 8). The median value of pH for the yearwas 5.1. Newman and Schalles (1990) report that for 49 bays that were studied theaverage pH was 4.6. The liming of the soils for agricultural production probably causedthe higher pH for Juniper Bay.

Precipitation

Table 5 compares the three bulk rain gauges and shows little spatial variation innutrient levels in the rainfall. The nutrient levels in the precipitation are generally lowerthan the amounts of nutrients in the surface water showing that precipitation is not amajor source of nutrients entering the bay. In fact, the nutrient levels are lower than thenutrients in the surface outflow, representing the fact that there is leaching of nutrientsfrom the soil into the surface water. The only nutrients that are higher in the rainfall thanthe surface outflow are NH4

+-N, NO3--N, PO4

3-, and TKN. This suggests that there mightbe some sort of deposition of these nutrients from outside sources into the bay. The factthat these higher levels of NH4

+-N, NO3--N, PO4

3-, and TKN do not make it to the surfacewater could result from the fact that the nutrients are either assimilated by plants orimmobilized in the soil.

193

Ground Water

Groundwater was sampled from possible sites of groundwater input and outputfrom the bay. Table 6 lists the average nutrient levels for each piezometer for the year.Concentrations of NH4

+-N, NO3--N, PO4

3- in the groundwater were low. However, inpiezometer 17 there were high levels of calcium. Cores collected at the site showed thata shell layer occurred at this location at a depth of 7 to 8 m. Piezometer 1 had high levelsof TOC, which might be caused by buried organic materials, which are also known tooccur at the site. According to core descriptions taken at this location at around 10 ftthere was evidence of charcoal and wood fragments, which is the depth of the watersamples.

The occurrence of high levels of nutrients at locations where it is thought that thereis groundwater inflow could indicate that there is groundwater movement into JuniperBay. For most nutrients, concentrations from all of the piezometers were either two tothree times higher than the nutrient levels in the surface outflow. The higher levels atthese locations could be diluted by the inflow of water, which would then show up in thesurface outflow having lower nutrient levels. Schalles and Shure (1989) indicate that thedilute chemistry of surface waters in Carolina Bays could suggest that subsurfacehydrologic exchange must occur.

The pH recorded for all of these locations in general is higher than the pH of thesurface outflow (Fig. 9). This is due to the high amount of cations in the groundwater atthese locations. Piezometer 17 particularly had the highest levels of calcium and also hadthe highest pH. Piezometer 3 has the lowest pH and has the lowest amount of calcium insamples.

Implications

On average the nutrient levels of P and N in the surface water were not a concernfor water quality degradation and eutrophication. Base flow values for NH4

+-N were 0.26mg/L, NO3

--N were below detection limits, TKN values were 1.21 mg/L, and PO43- were

0.02 mg/L. Outflow values for storm events less than 5 cm a day had nutrient levelssimilar to those of the base flow. Rain events larger than 5 cm a day increased theconcentrations of nutrients by two to three times the base flow conditions, and did raisethe levels to where there might be short term algae blooms. However, these raisedphosphorus levels were low enough not to cause long-term eutrophication. On averagethe level of P in the surface water was within the natural level limits of 0.05 to 0.05mg/L. Since the phosphorus concentrations are so low, there should not be any excessalgae blooms that would lead to water quality degradation.

Sampling surface outflow by looking at the differences in nutrient levels duringstorm events showed that only large (> 5 cm a day) storm events needed to be sampledseparately. These two large storms that occurred accounted for over two thirds of thetotal mass of nutrients exported from the bay (Table 7). This indicates that large stormevents are a significant source of nutrient export and is important to measure these

194

events. Sampling during periods of baseflow condition could be done less frequent andstill get good results.

Juniper Bay on average during base flow conditions has almost three times theamount of calcium in its surface water than those bays that were sampled by Newmanand Schalles (1990). This higher level in the surface outflow could be a result of the limeand fertilizer applications that occurred at Juniper Bay during agricultural production.Since the halt of agricultural production, the nutrients are being leached out of the soiland are leaving the bay in surface water.

It seems that there is a steady decrease in nutrient levels and pH in Juniper Baysince the end of agricultural production in 2000. The pH of Juniper Bay is not knownduring agricultural production, but it is thought that is was higher than the currentaverage pH of 5.1. Since the application of nutrients has stopped, the cations are leachingout of the soil allowing for the pH to return to natural levels.

Schafale and Weakly (1990) report that Carolina Bays are wet and nutrient poorcommunities. It seems that with the lack of nutrient inputs from precipitation, JuniperBay will become more nutrient deficient in time and will resort back to natural baynutrient levels. This loss of nutrients is a good indication that Juniper Bay is in theprocess of returning to a nutrient poor community. The loss of nutrients will mean thatplants that are native and adapted to such sites will have a good chance of beingreestablished as part of the mitigation project.

CONCLUSIONS

Nutrient analyses were performed on the water entering and leaving adrained Carolina Bay in order to determine the nutrient input and outputs, and todetermine if there was the potential to cause water quality degradation. Our findingsindicate that precipitation is not a major source of nutrient input into the bay. Thegroundwater nutrient analysis revealed higher levels than in the surface water indicating apossible dilution of nutrients by groundwater inflow into ditches.

Surface outflow nutrient levels on average where quite low with littleconcern that the water leaving Juniper Bay will cause eutrophication at this time. Therewas little difference between base flow conditions and rain events less than 5 cm a day.However, for rain events greater than 5 cm a day there was a two-fold increase in nutrientconcentration in surface outflow. This increase can be attributed to the heavy rain thatexceeded the infiltration rate of the soil and caused surface runoff. These nutrient levelsstay elevated for about 1 week because of the time it takes for the whole bay to drain.There could possibly be short-term water quality degradation with these storm events butlong-term eutrophication should not be a problem.

The sampling scheme at Juniper Bay could be such that surface water samples aretaken less frequent, except during large storm events. These large storm events even

195

though they were few and infrequent, still accounted for a significant amount of the totalexport of nutrients from the bay and are an important component to measure.

With the loss of nutrients from Juniper Bay, the natural nutrient conditions arereturning to this site. This return to a nutrient poor community is indicates that plantsthat are adapted to these conditions will be able to reestablish at the site and allow forJuniper Bay to be restored back to its original community.

196

REFERENCES

Alberta Natural Heritage Information Centre. 2002. Alberta Lentic Wetland HealthAssessment (Survey) User Manual. Available at: www.cowsandfish.org/pdfs/AlbertaLenticSurveyUsersManual.pdf.

Bormann, F.H., and G.E. Likens, 1967, Nutrient Cycling, Science, Vol 155, Issue 3761pg 424-429.

Carter, V. 1986, An overview of the hydrologic concerns related to wetlands in theUnited States. Can. J. Bot. Vol 64 pg 364-374.

Carter, V., M.S. Bedinger, R. P. Novitzki, and W. O. Wilen. 1978. Water resources andwetlands In Wetland Functions and Values: The State of Our Understanding, P.E.Greeson, J.R. Clark, and J.E. Clark eds., American Water Resources Assoc.,Minneapolis, Minn., p. 344-376.

Dodds, Walter, K., and Eugene B. Welch, 2000, Establishing nutrient criteria in streams,J.N. Am. Benthol. Soc., Vol 19(1) pg 186-196

Hauser, J. NCDOT Develops Juniper Bay Mitigation Site. Centerline An EnvironmentalNews Quarterly, From the NCDOT Natural Systems Unit. January 2001 IssueNo. 4.

Johnson, P.L., and W.T. Swank. 1973. Studies of Cation Budgets in the SouthernAppalachians on Four Experimental Watersheds with Contrasting Vegetation.Ecology, Vol 54. No. 1 pg. 69-80.

Likens, G. E., F. H. Bormann, N. M. Johnson, and R. S. Pierce. 1967. The calcium,magnesium, potassium and sodium budgets for a small forested ecosystem. Ecology 48: 772-785.

Newman, Michael, C. and John F. Schalles, 1990, The water chemistry of Carolina bays:a regional survey, Arch. Hydrobiol. Vol 118, pg 147-168.

Osmond, D.L., D.E. Line, J.A. Gale, R.W. Gannon, C.B. Knott, K.A. Bartenhagen, M.H.Turner, S.W. Coffey, J. Spooner, J. Wells, J.C. Walker, L.L. Hargrove, M.A.Foster, P.D. Robillard, and D.W. Lehning. 1995. WATERSHEDSS:Water, Soiland Hydro-Environmental Decision Support System,http://h2osparc.wq.ncsu.edu.

Radtke, D. B. 1999 United States Geological Survey. . Processing of Water Samples-Sample Preservation Chapter 5.4.2. Available at:http://water.usgs.gov/owq/FieldManual/chapter5/html/5.4.2.html

Schafale, M. P. and A. S. Weakly. 1990. Classification of the Natural Communities ofNorth Carolina Third Approximation. North Carolina Natural Heritage Program

197

Division of Parks and Recreation Department of Environment, Health, andNatural Resources P.O. Box 27687 Raleigh, NC.

Schalles, J. F., and D. J. Shure. 1989. Hydrology, Community Structure, andProductivity Patterns of a Dystrophic Carolina Bay Wetland. EcologicalMonographs 59(4) 365-385.

Sharitz, R.R., and J. W. Gibbons. 1982. The Ecology of Southeastern Shrub bogs (Pocosins) and Carolina Bays: A Community Profile.

USEPA. (US Environmental Protection Agency). 1986. Ambient Water Quality forBacteria-1986. EPA 440/5-84-002.

USEPA. (US Environmental Protection Agency). 1998. Notice of availability of cleanwater action plan. Federal Register 63:14109-14112.

198

PO4 NH4 NO3 TKN TOC Ca K Mg Na

Base flow (mg/L)

Day 45-50 0.02 0.00 0.00 1.45 44.00

Day 54-59 0.01 0.00 0.00 0.67 44.30

Day 142-147 0.02 0.17 0.00 1.07 27.50 4.90 4.05 2.42

Day 166-170 0.02 0.44 0.00 1.25 19.50 4.55 3.83 2.25

Day 181-183 0.03 0.25 0.00 0.84 14.00 4.00 3.15 1.90 3.40

Day 240 0.01 0.69 0.00 2.00 21.00 4.40 4.40 2.00 5.30

Mean (6) 0.02 0.26 0.00 1.21 28.38 4.46 3.86 2.14 4.35

PO4 NH4 NO3 TKN TOC Ca K Mg NaStormflow (mg/L)

<5 cm

Day 173-174 0.02 0.38 0.00 1.08 15.00 4.25 3.65 2.05 3.45

Day 165 0.02 0.41 0.00 0.98 12.00 3.20 2.60 1.60 3.20

Day 121-128 0.02 0.26 0.00 1.06 40.00

Mean (3) 0.02 0.35 0.00 1.04 22.33 3.73 3.13 1.83 3.33

Table 1. Nutrient concentrations in surface outflow for several periods during theyear with no rainfall (baseflow) conditions

Table 2. Nutrient concentrations in surface outflow for several periods duringthe year that there was rainfall less than 5 cm in a day.

199

PO4 NH4 NO3 TKN TOC Ca K Mg Na

Stormflow > 5 cm (mg/L)

Day 241 0.05 0.12 0.13 2.70 85.00 12.40 9.30 5.30 3.00

Day 285 0.13 0.23 0.27 2.2 74 17 8 6.2 3.00

Mean (2) 0.09 0.18 0.20 2.45 79.50 14.70 8.65 5.75 3.00

Season PO4 NH4 NO3 TKN TOC Ca K Mg Na

(mg/L)

Winter 0.02 0.01 0.04 1.44 40.23 9.53 7.37 5.97 3.67

Spring 0.02 0.09 0.00 0.91 39.04 5.08 4.93 2.54 3.16

Summer 0.03 0.33 0.00 1.47 30.95 5.86 4.41 2.70 3.54

Fall 0.04 0.40 0.06 1.87 50.24 8.96 6.37 4.47 4.28

Table 3. Nutrient concentrations in surface outflow for the two rain eventsthat were greater than 5 cm of rain in a day.

Table 4. Average nutrient concentrations based on season at Juniper Bayin outflow water.

200

PO4 NH4 NO3 TKN TOC Ca K Mg Na(mg/L)

NW 0.23 1.83 0.12 2.90 7.62 0.82 1.12 0.28 1.82SE 0.30 1.88 0.15 3.87 5.70 0.98 1.45 0.35 2.35WS 0.16 1.36 0.09 2.68 6.43 0.77 0.87 0.22 1.45

PO4 NH4 NO3 TKN TOC Ca K Mg NaWellno.

(mg/L)

inflow 1 0.07 0.50 0.00 5.48 155.20 40.78 16.74 2.48 11.08outflow 3 0.02 0.49 0.02 1.84 35.60 9.63 7.35 1.95 6.93inflow 17 0.00 0.22 0.06 0.94 35.33 99.70 3.70 2.10 6.18outflow 25 0.32 0.59 0.04 2.33 63.40 17.53 15.93 5.78 12.18

Table 5. Average nutrient concentrations for bulk precipitation gauges located inJuniper Bay

Table 6. Average nutrient concentrations for groundwater for four locations in JuniperBay.

201

Table 7. Comparison of storm events and mass export of nutrients out of bay.The two large storm events accounted for a significant amount of nutrient export.

Stormsize

PO4 NH4 NO3 TKN TOC Ca K Mg Na

cm/day (kg/yr)< 5 4.00 52.00 0.00 242.00 5676.00 892.00 772.00 428.00 870.00

> 5 10.62 21.24 23.60 289.10 9381.00 1734.60 1020.70 678.50 354.00

TOTAL(kg/yr)

14.62 73.24 23.60 531.10 15057.00 2626.60 1792.70 1106.50 1224.00

> 5cmas % oftotal

73 29 100 54 62 66 57 61 29

202

Figure 1. Map of Juniper Bay showing locations of bulk precipitationcollectors.

Bulk precipitationcollectors

203

Figure 2: Schematic diagram of Bulk Precipitation Collector(Modeled after Liken et al. 1967).

Polyethylenefunnel 8-inchdiameter

VaporBarrier

Polyethylenereservoir (2 liters)

Tygon Tubing8 ft 4x4wooden post

204

outflow

0

50

100

150

200

250

300

350

400

136 137 138 139 140 141 142 143 144

day of year

cubi

c fe

et

Bottles 1-4 composite

Bottles 9-11 composite

Bottles 12-15 composite

Bottles 5-8 composite

Bottles 16-21 composite

Figure 3. Graph of surface outflow and bottle numbers. This method was used tocomposite sample bottles together that represented similar periods in the hydrograph.

205

1

2517

3

Figure 4. Juniper Bay showing location of where groundwater was sampled.Numbers represent piezometers that where used for collection of groundwater.

206

0

1000

2000

3000

4000

5000

6000

239 241 243 245 247 249

day of year

cub

ic f

eet

0

10

20

30

40

50

60

70

80

90

100

TO

C m

g/L

outf lowTOC

Rain event

Figure 5. Total organic C concentrations during a rain event greater than 5 cm in aday. Increase of nutrient occurred a couple of days later and nutrient levelsremained high for a week.

207

0

1000

2000

3000

4000

5000

6000

239 241 243 245 247 249

day of year

cub

ic f

eet

0

2

4

6

8

10

12

14

mg/

L outflow

Ca

Rain event

Figure 6. Calcium concentrations during rain event greater than 5 cm in a day. Calciumpeaks a couple of days after rain event and takes a week for level to almost return to baseflow conditions.

208

Figure 7. Juniper Bay showing possible drainage regions, which could be apossible explanation as to why it takes so long for nutrient levels to decreaseafter large rain event. The areas closest to the main outflow point would drainthe fastest. The area shaded in red is farther away and would take about a weekto drain.

2 days

4 days

6 days

10 days

209

0

1

2

3

4

5

6

7

2/1/2002 3/23/2002 5/12/2002 7/1/2002 8/20/2002 10/9/2002 11/28/2002 1/17/2003

pH

Figure 8. Weekly pH readings for surface water at main outflow. The average pH for theyear is 5.1

210

pH

0

1

2

3

4

5

6

7

8

Mar-02 May-02 Jun-02 Aug-02 Oct-02 Nov-02 Jan-03 Mar-03

date

pH

1

3

17

25

Figure 9. pH measurements for groundwater at different locations in Juniper Bay. 17 had thehighest pH and had the highest amounts of Calcium. 3 had the lowest pH and had the lowestamounts of Calcium.

211

Chapter 9

GROUND PENETRATING RADAR EVALUATION OF JUNIPER BAY’SSUBSURFACE STRATIGRAPHY

R.P. Szuch, J.G. White, and M.J. Vepraskas

INTRODUCTION

Clayey soil horizons within Juniper Bay should act as aquitards and may aid thehydrologic restoration of the site. Where such aquitards are near enough to the surface,they may foster inundation or saturation such that wetland hydrology requirements willbe met. The presence of such aquitards also might aid the establishment of hydrophyticvegetation. Where aquitards are not present, precipitation should infiltrate to the watertable and have no appreciable impact on surface or near-surface hydrology. The depth,extent, and continuity of clayey soil horizons are being investigated via ground-penetrating radar (GPR).

MATERIALS AND METHODS

The GPR technique allowed spatially continuous transects of the site to besurveyed for the presence of clayey soil horizons. To date, GPR surveys have beenconducted in 15 "fieldlets" at Juniper Bay (Fig. 1), totaling 23.2 km of GPR images.Three GPR transects were performed in each fieldlet: (1-center) longitudinally down thecenter, (2-edge) longitudinally along a ditch edge, (3-cross) laterally across the fieldlet.

Figure 1. Map of Juniper Bay and ditch network."Fieldlets" (areas enclosed by ditches) where GPRsurveys have been performed are highlighted.

All GPR fieldwork and data post-processing were supervised by Jim Doolittle ofthe NRCS. The GPR unit is dragged along the soil surface as it emits high-frequencyelectro-magnetic waves. These waves penetrate the soil, and some of the energy is

212

reflected when a stratigraphic interface (such as sandy to clayey soil) is encountered. TheGPR unit measures the travel time for the waves to reach the stratigraphic interface andreturn to the surface. Associated hardware and software produce an image of the GPRreturn. The images have an x-dimension of distance along the ground surface and a y-dimension of two-way travel time (time for GPR waves to travel down and reflect back).Clayey soil horizons manifest themselves as bright reflections along the GPR images(Figure 2).

Figure 2. Typical GPR image, showing a bright reflection delineated with athick black line. The x-dimension is distance along the ground surface, andthe y-dimension is two-way travel time in nanoseconds.

Using a scalebar (as in Fig. 2), the location of bright reflections were interpreted from theGPR images. However, the interpretations had to be calibrated so that the y-dimension oftwo-travel time was converted to a useable depth value. This involved two steps. First,we determined the location of the soil surface on GPR images. Second, we developed acalibration equation to convert the y-dimension from time to depth.

The top of GPR profiles contains several straight, parallel lines. These lines donot represent the soil surface or any subsurface interface. In the literature, the line thatrepresents the soil surface is not always clearly delineated. A simple method todetermine the location of the soil surface was to begin a GPR scan with the antennastationary on the ground, lift it to shoulder-height (Fig. 3a), and return it to the soilsurface. Relative to the antenna, the soil surface was deeper when the antenna wasraised; this resulted in a dip on the GPR profile. The uppermost dipping line wasidentified as the soil surface (Figure 3b).

0ns

ns

50

100

150

200

2

7

Distance Along Ground Surface

213

Figure 3. (a) Demonstration of “lift method” at JB. (b) Resulting GPR trace, soilsurface shown with dotted black line.

The calibration equation was determined using reflector-interface matching. Thedepth of a clayey-sandy interface was determined from soil coring, and the interface wasmatched to a bright reflector on a GPR trace at the coring site. This calibration methodyields a time-depth data pair: time - drawn from the GPR’s time scale, and depth - drawnfrom the core’s depth scale. This method is relatively accurate, especially when thecoring is done immediately after the GPR survey. A reflector-interface pair was obtainedat 16 locations throughout Juniper Bay. This calibration data was plotted on a scatterdiagram and a linear trendline and equation were obtained (Fig. 4). Velocity of GPRwaves at JB was highly dependent on saturation due to the water table (see Table 1).When the interface was beneath the water, the velocity was nearly constant and wasconsiderably slower than when the interface was above the water table. On days of GPRsurveys, the water table was above most clayey-sandy interfaces that were of interest tothis study. The “below water table” calibration points were used to obtain a linearequation between two-way travel time to a reflector and depth to a stratigraphic interface(see Figure 4: y = 0.0274x + 0.1631, R2 = 0.8691). This equation was used to convert theinterpretations from GPR surveys, which are obtained in time (ns), to depths (m).

(a)

(b)

214

Figure 4. Relationship between depth to clayey-sandy interface andtwo-way travel time to reflector (n=16), with equations of trendlines andR2. The data is split based on the interfaces’ relation to the watertable. The equation to the right would be used during interpretation ofGPR surveys.

Table 1. Average, maximum, and minimum GPR wave velocities based on calibrationpoints above and below the water table.

Dataset Avg. Velocity (m/ns) Max. Vel. Min. Vel.Above Water Table 0.115 0.133 0.099Below Water Table 0.063 0.079 0.052

Coring was performed during the summers of 2002 and 2003 to ground-truth theGPR interpretations. Most coring was performed by hand-augering to a depth of 1-3 m.The actual depth of clayey layers found via coring was compared with the depths aspredicted via GPR.

RESULTS AND DISCUSSION

Actual and predicted depth to clayey layers were compared by plotting both onExcel diagrams (Fig 5). This has been performed for all "cross" and "center" GPRsurveys within each fieldlet. To date, accuracy has ranged from 1 to 95 cm and averages25 cm. This average accuracy is considered good for this type of geophysical survey.

y = 0.0274x + 0.1631

R2 = 0.8691

y = 0.0392x + 0.2386

R2 = 0.7968

0

0.5

1

1.5

2

2.5

0 10 20 30 40 50 60 70

Two Way Travel Time to Reflector (ns)

below water table

above water table

Linear (below water table)

Linear (above water table)

215

The spatial variation in accuracy can be viewed in Figure 6. Further investigation isbeing performed to see if interpretations can be improved in regions of the bay with pooraccuracy. Ground-truthing along the "edge" GPR surveys will be concluded in the latesummer of 2003.

216

Figure 5. Graph showing depth of clayey horizons as determined via coring and viaGPR interpretations. Vertical lines indicate locations where coring was performed.Colors indicate the soil texture found. Colored horizontal lines indicate the location ofclayey horizons according to GPR interpretations. Accuracy is measured asdifference between the horizontal lines and the point where sandy soil changes toSCL, SC, or clay(C).

Figure 6. Map of Juniper Bay showing spatial variation in accuracy. Accuracy being theabsolute value of the difference between clayey horizon depth as determined by coring versusGPR interpretation.

INTERFACES OF JI 03

0

0.5

1

1.5

2

2.5

3

0 100 200 300 400 500 600

Distance along Trace (m)

Dep

th to

Inte

rfac

e (m

)

SCSCLsandy clay

217

Once the calibration and accuracy of the GPR survey was established, we beganto address our general aim - to investigate the depth, extent, and continuity of clayey soilhorizons via GPR. We determined that this aim could be expressed as three distinctobjectives:1- Find areas where aquitards are not present or very deep2- Find ditch-induced aquitard discontinuities3- Find natural aquitard discontinuities

GPR interpretations show that aquitard depth ranges from 0.5 to 3.0 m with anaverage of 1.5 m. One, two, or more aquitards are present throughout Juniper Bay, withaquitards frequently overlying each other. However, the GPR shows an anomaly in thesoutheast corner of Juniper Bay that may indicate a region with no aquitard present.Coring in this region has confirmed that no clayey soil horizons exist near the surface.The predicted extent of this anomaly, based on GPR surveys, is shown in Figure 7. Thesize of the anomaly is estimated to be 3 ha, or approximately 1% the area of Juniper Bay.No aquitard exists in this anomaly, or the aquitard is at a depth below current coringdepth (deepest have been nearly 5 m). In either case, surface and near-surface hydrologyshould not be impacted by any clayey soil horizons in this region.

Figure 7. Map of Juniper Bay showing the location of an anomaly (redband) in the GPR survey that seems to have no clayey aquitard near thesurface.

218

Ditch induced aquitard discontinuities would result where the depth of a drainageditch is greater than the depth of a clayey soil horizon. A GIS coverage of Juniper Bay'stopography, provided by NCDOT, was used to determine the depth of ditches throughoutthe bay. This data was compared with the depth of clayey aquitards as determined by theGPR interpretations. According to this comparison, ditches have pierced aquitards in atleast 18 locations throughout the bay, predominantly along main N-S ditch. Coring hadbeen done near some of these locations, and coring data confirmed that ditches shouldhave pierced the aquitards in 10 of the 18 locations. Thus far, this analysis has only beenperformed with the "center" and "cross" GPR surveys. Comparing the ditch depths to the"edge" survey should yield more valuable results.

Natural aquitard discontinuities might be found in the interior of fieldlets, perhapswhere two adjacent clayey horizons do not overlap and water could infiltrate below thesepotential aquitards. There is some evidence of natural discontinuities in several regionsof the bay, but these occurrences require further investigation and documentation beforewe are prepared to report on them.

CONCLUSIONS AND RECOMMENDATIONS

GPR surveys have been conducted successfully in the interior of a drainedCarolina Bay. To our knowledge, this had not been previously accomplished. The “liftmethod” for defining the soil surface on GPR traces should increase accuracy ofinterpretations. Calibration via reflector-interface matching has proven an effectivemethod at JB. A calibration equation based on multiple calibration points, from variouslocations and depths, was created (Fig. 4). This approach appears to be advantageous fora site of such great size and complexity. Interpretation of GPR surveys from JB iscomplete, and its application toward the stated objectives is continuing. At the currenttime, two important findings are brought to the attention of NCDOT. First, an anomalyof approximately 3 ha exists in the southeast corner of Juniper Bay. Unlike the rest ofJuniper Bay, this anomaly appears to have no clayey aquitards. NCDOT may want toconsider the ramifications of this finding in their restoration plans. Second, it seemslikely that at least the main N-S ditch has pierced underlying aquitards. Any water thatmight have been retained near the surface by these aquitards will be drained off site.NCDOT may want to consider a clay or synthetic lining in the base of the main ditch orother ditches to prevent such water loss. More detailed and complete informationregarding ditch-induced aquitard discontinuities should be available in fall of 2003.Information on natural aquitard discontinuities should be available in fall of 2003 as well;however, due to the location of such discontinuities in the interior of fieldlets, it does notseem likely that NCDOT would take any action to account for their presence.

219

Chapter 10

GROUNDWATER HYDROLOGY OF JUNIPER BAY PRIOR TORESTORATION, AND GROUNDWATER HYDROLOGY OF REFERENCE

BAYS

S. Luginbuhl, J.D. Gregory, and M.J. Vepraskas

INTRODUCTION

This portion of the overall project evaluated the pre-restoration ground waterhydrology at Juniper Bay. Major objectives were to determine the current ground waterflow paths and the water table regime both inside and outside the bay. In addition, thewater table fluctuations in the three reference bays were monitored to determine whetherparts or all of the reference bays met wetland hydrology requirements. These data wouldbe useful in planning the restoration strategy.

MATERIALS AND METHODS

A network of shallow water table monitoring wells, 28 total, was installed insideand outside the bay to determine that the water table regime at various locations in thebay and to determine lateral hydraulic gradients in the near surface ground water (Figure1). A network of piezometers, nested with the water table monitoring wells was installedinside and outside the bay to determine vertical and lateral hydraulic gradients below thenear surface clay layers in a plan to collect pre-restoration data for two years and post-restoration data for three to five years. There are a total of 11 nests of piezometers andwells (Figure 1). Figures 4 through 7 include water table hydrographs for all Juniper Bayand reference bay wells. Appendix A includes all graphs from the nests of piezometers.Another goal was to determine how the water table regime is affected by rainfall, soilcharacteristics, topographic variation, and near-surface hydraulic gradients. Figure 2shows core locations.

In each reference bay, four monitoring wells were installed in one transect fromthe rim to the center of the bay. The first well was installed in mineral soil, the secondwell in mineral soil with a histic epipedon, the third well in shallow organic soil, and thefourth well in deep organic soil. The goal was to determine how the water table regimewas affected by rainfall, soil characteristics, and vegetation type. Three bays wereselected as the reference bays to provide data on the variability among natural CarolinaBays.

220

RESULTS AND DISCUSSION

Juniper Bay

Some of the main factors that affect the hydroperiod of Juniper Bay and thereference bays are rainfall, soil type, elevation, and vegetation. Carolina Bays typicallyhave hydrologic inputs consisting mainly of precipitation and, as some suggest, groundwater discharge (Lide et al, 1995; Schalles and Shure, 1989). Some bays experiencesurface run on as another input, but Juniper Bay does not have surface run on due to theperimeter ditch. The hydrologic outputs in Juniper Bay consist mainly ofevapotranspiration, surface outflow, and possible ground water recharge. Soil typeinfluences how much rainfall infiltrates and percolates down through the soil profile.Elevation influences direction of surface and ground water flow, and the vegetationpresent on the site influences how much soil water in lost through evapotranspiration(ET) and how much water reaches the soil surface. A dense canopy cover will reduce theamount of water that enters the soil, as in the reference bays.

Rainfall

The monthly, seasonal, and annual averages from 1971 through 2000 for rainfallat the weather station in Lumberton, North Carolina and Elizabethtown, North Carolina(N.O.A.A., 2001), and from the weather station at Juniper Bay are shown in Table 1.Juniper Bay is about 12 kilometers south of Lumberton, Tatum Millpond Bay is about 12km north of Elizabethtown, Charlie Long Millpond Bay is about 17 km north, andCauseway Bay is about 18 km east northeast of Elizabethtown.

As Table 1 shows, the yearly averages for Juniper Bay in 2001 and 2002 arebelow the long term averages. Certain months and seasons, however, were abovenormal. The summer of 2001 was wetter than average, as was the fall of 2002. The fallof 2001, though, was far drier than normal. Fall is typically the time before the wetseason, and normally has the least amount of rain. Summer usually has the most rainfall,followed by winter, than spring. The water table data (Figs. 4 to7) shows that many wellsshowed a drop, some sharp, in the water level during the fall of 2001. Most also show asignificant drop in the summer of 2002. This second drop was likely due to increased ETdue to the heat of the summer and the denser, taller vegetation that was present on the sitecompared with last summer. The summer of 2002 was also hotter than the summer of2001 and had slightly below average rainfall.

Juniper Bay Water Table Regime

A wetland is defined by its hydrology, soils, and vegetation. The hydrologycomponent varies slightly between agencies, but generally it is that the water table muststay within 30 cm of the soil surface for a certain amount of time during the growingseason in most years. The amount of time is important because water must be

221

consistently present at the 30 cm or shallower depth for a certain amount of time for thesoils to become hydric and for hydrophytic plants to dominate. The hydrologycomponent is often said to be the most important component because if it is not present,hydric soils and hydrophytic plants will not exist for long.

Monitoring the hydrology of Juniper Bay before restoration is important in orderto determine which areas of the bay may pose more problems when trying to restorewetland hydrology. There were 28 water table monitoring wells in and around the bay(Fig. 1). Parts of the bay already met some of the hydrology requirements (mainly thesoutheast section). The percentage of Juniper Bay that meets wetland hydrologyrequirements depends on which requirements are used. According to Corps of Engineerscriteria, 22% of the bay meets wetland hydrology requirements, which are that the watertable be within 30 cm of the soil surface for at least 5% of the growing season in mostyears (Table 4). The Corps uses 5% as the criteria for wetland delineation purposes and12.5% as the criteria for wetland restoration purposes. Only 12% of the bay meets thecriteria for wetland restoration.

Response of Water Table to Precipitation Events

Well 2A is located in mineral soil near the rim, in the northwest section of JuniperBay. Its elevation is the second highest of the wells in that section, at 36.47 m, and thatsection is the highest in elevation in the bay. In the last days of May, 2001, it rainedabout 0.4 cm. The water table in well 2A did not begin to rise until the third day of rain.On the 151st and 152nd day of 2001 (the beginning of June), there was an additional 0.43cm of rainfall, for a total of 0.83 cm of rain over 7 days. The water table roseapproximately 61 cm in those 7 days. For two days in the middle there was no rain, andthe water table had a brief period (day 152) in which it barely rose (Fig. 4). The watertable rose from day 149 to day 153 even though the rain events took place on day 147through 149 and days 151 and 152. Another rain event took place on days 230 through233. The rainfall totals were approximately 1.23 cm. The water table rose almost 59 cmfrom day 232 to 233. The above rainfall totals and subsequent water table rise indicatethat in the last spring of 2001 a rain event that produced approximately 0.83 cm of rainresulted in a water table rise of about 61 cm, while a larger rain event of 1.23 cm in lateAugust of 2001 resulted in a slightly smaller but basically equal water table rise of 59 cm.The additional rain that fell in August was likely eliminated by the increased rate of ET inthe late summer, resulting in a relatively equal rise in the water table. The other wellsbehaved in a similar manner. A day or two after a rain event the water table would startto rise and would stop rising a day or two after the rain event ceased (Figs. 4to7).

Water table well 2A showed seasonal variation. During the summer of 2001, thewater table was 80 cm below the surface for the entire season. Summer 2001 did havegreater rainfall than the long term averages, which likely accounted for the higher watertable. In the fall of 2001, a very dry fall (Table 1), the water table dropped steadily about40 cm. During the winter of 2001/2002, the water table recovered, with the help of adecent amount of rain (still below the long term average), and lower temperatures. Thespring of 2002 saw the water table spike up and down, and then steadily drop down (with

222

a few spikes due to rain events) towards the summer of 2002. The summer of 2002 hadbelow average rainfall and the water table dropped to a low of -142 cm below the soilsurface in mid-August. By the fall of 2002 the rainfall became closer to the average andthe water table rose again. This seasonal pattern was evident in all water table wells inJuniper Bay. All wells saw water table declines during the fall of 2001 and the summer2002 due to lack of rainfall.

Water table wells 8, 12, and 16 (Fig. 1) showed the water table staying above orjust under the soil surface for much of the winter and spring of 2002 (see Fig. 6). Well 8is located in organic soil but wells 12 and 16 are not. The soil profile at well 8 consists ofa muck layer over a thin sand layer, which is over a very thick clay layer. Water at thislocation cannot percolate easily down through the soil, and therefore stays closer to thesurface during the wetter months. During the dry summer of 2002, however, the watertable dropped very deep. The affects of lack of rain, thick clay and hot temperaturesmade water at this location scarcer than at other locations in the bay. Wells 12 and 16have similar soil profiles that consist of a thin muck layer at or near the surface and overmainly sand, with thin clay lenses also present.

Water table levels in Juniper Bay rise with rainfall totals as small as 0.762 mmand then fall once the rain slows to 0.254 mm or less. The fall of the water table is mostlikely due to the ditches plus some ET and evaporation of water from the top layer of thesoil.

The soil stratigraphy can greatly affect water table movement. Some of the areasin the bay have thick clay layers near the surface, which may perch water on top of themcausing that area to be wet (such as the area around well 8). Other areas have thick sandlayers which allow surface water to drain away. There is a wide variety of soil profiles atJuniper Bay, which adds to the complexity of its hydrology. The wells that meet wetlandhydrology requirements, 12.5% of the growing season, have varying soil profiles. Thelocation of well 8 has a muck layer about 45 cm thick and is an organic soil. Under themuck is clay. Wells 12 and 16 are in mineral soils with muck layers near the surface.They are all located in a section of the bay where elevations are lower compared with theother sections of the bay. Other areas of the bay with similar profiles do not meetwetland hydrology requirements, so it has something to do with this area of the bay,possibly the elevation, or a lack of crowning of the fields in this area.

The surrounding land use may play a part in whether the bay will exhibit “typical”bay hydrology once restoration is complete. Since most of the land adjacent to JuniperBay is in agriculture, the regional water table may be lower than it was years ago. It maybe unrealistic to try to get the water table in Juniper Bay to behave as it did prior todrainage. The regional water table around Juniper Bay may be lower than the regionalwater table around the reference bays due to this agricultural land use. The land aroundat least two of the reference bays is not in heavy agriculture. This may allow thereference bays to be wetter than Juniper Bay could ever be.

223

Reference Bays

Rainfall patterns in the reference bays are similar to those in Juniper Bay (Table1). The water table data from the wells in the reference bays indicate that there is morewater present in these systems. The water table is consistently higher in the mineral soilsin the reference bays than in the mineral soils in Juniper Bay, and this same pattern existsin the shallow organic soils and in the transition soils (mineral soils with histicepipedons). Vegetation in the reference bays is much denser than in Juniper Bay,consisting of trees and shrubs whereas the vegetation in Juniper Bay consists of mainlyweeds. This would mean that ET is higher in the reference bays and that less rainfall isreaching the surface due to interception by the leaves. Yet the water table is closer to thesurface, with most wells exceeding all wetland hydrology requirements. The wells in theorganic soils had water tables closer to the surface, followed by wells in the mineral soilwith a histic epipedon, and the wells in the mineral soil had water tables that were thedeepest. Even though the water tables in the mineral soils in the reference bays weregenerally shallower than those in Juniper Bay, the water table in Juniper Bay spikedabove the water table in the reference bays. This did not happen very often in the othersoil types; the water table in the reference bays generally remained shallower than thewater table in Juniper Bay. The only wells in Juniper Bay that spiked above the watertables in the reference bays were the wells that had water tables shallow enough to meetwetland hydrology requirements.

U.S. Army Corps of Engineers requirement for wetland hydrology is inundationor saturation to the surface continuously for 5% of the growing season in most years(50% probability of reoccurrence) (U.S. Army Corps of Engineers, 1987). Requirementfor this restoration project by the North Carolina Department of Transportation is 12.5%of growing season in order to be closer to the hydroperiod of the reference bays. We willalso use 8.75% of the growing season to assess wetland hydrology.

As shown in Table 5, 22% of Juniper Bay has a water table shallow enough tomeet the minimum wetland hydrology requirement of 5% of the growing season, as setby the Army Corps of Engineers. As the time requirement for number of consecutivedays the water table has to be within 30 cm of the soil surface increases, the percentage ofwells in Juniper Bay that meet the requirement decreases. When the requirement is12.5% of the growing season, only 12% of the bay is wet enough.

Figure 7 shows the data from the wells in the three reference bays. U.S. ArmyCorps of Engineers requirement is inundation or saturation to the surface continuously for5% of the growing season in most years (50% probability of reoccurrence) (U.S. ArmyCorps of Engineers, 1987). Requirement for this restoration project by the NorthCarolina Department of Transportation is 12.5% of growing season in order to be closerto the hydroperiod of the reference bays. I will also use 8.75% of the growing season toassess wetland hydrology.

224

Ground Water Movement

The soils in Juniper Bay are a mixture of mineral and organic, and are all hydricsoils. Where the soil transitions from mineral to organic it is a mineral soil with a histicepipedon (a thin layer of organic soil at the surface too thin to be let the soil be classifiedas organic). The organic soils occur mostly in the center of the bay, the mineral soilswith histic epipedons encircle the organic soils, and the mineral soils occur closest to therim of the bay. Beneath the surface at different depths, there is a clay layer, in varyingthickness throughout the bay and probably continuous, that acts as a restrictive layer towater movement. However, the large collector ditches may cut through this clay layer.This clay layer occurs between 0.3 to 4.27 m below the soil surface. The sandy ororganic material above this clay layer and below the soil surface we have called aquifer“zone 1” (Fig. 3). Below this restrictive layer of clayey material more sandy materialoccurs, in most areas, and this area of sandy material is called aquifer “zone 2”. Belowzone 2 is the top of the Black Creek Formation, a clay layer that acts as another restrictivelayer and the bottom of the bay. Beneath the top of the Black Creek Formation isalternating sand and clay material and is called “zone 3”. Within zone 2 there are thinclay layers in some locations.

The North Carolina Coastal Plain is an eastward-dipping and thickening sequenceof sand, silt, clay, and limestone. Beds primarily consisting of sand or limestonecompose aquifers, and beds largely consisting of clay and silt are confining units (Giese,Eimers, and Coble, 1992). Ground water follows the eastward-dipping geologic materialtowards the Atlantic Ocean. Ground water generally enters the confined aquifers ininterstream areas (Giese et al, 1992). Juniper Bay is located between the Lumber Riverand the Pee Dee River, but it is closer to the Lumber River. This means that water isentering the confined aquifer, zone 2 and zone 3, to the south of Juniper Bay, closer to theinterstream divide between the Lumber and Pee Dee Rivers. Ground water enters theconfined aquifers in the interstream divides because this is where there is the least pullfrom the large river systems. The geologic makeup of Juniper Bay consists of thesurficial aquifer, the Black Creek confining unit, and then the Black Creek Aquiferunderneath. Regional water table contours for the surficial aquifer (A10) show that thewater table flows towards the Atlantic Ocean in general, unless a large river is nearby, inwhich case the water table flows towards the river (Giese et al, 1992). The same is truefor ground water above the Black Creek confining unit and in the Black Creek Aquifer(possible zone 2 and zone 3 ground water in Juniper Bay, respectively).

Hydraulic Gradients

Hydraulic gradient is defined as the change in total head with a change in distancein a given direction. The direction is that which yields a maximum rate of decrease inhead; in other words, ground water moves from regions of higher head to regions oflower head (Fetter, 2001). The localized ground water flow patterns within Juniper Bayare complicated. In the zone 2 region, ground water seems to flow towards the center ofthe bay from the northwest and southeast ends of the bay, and then out of the bay throughthe north and south sides. This follows the localized topography, with the west and

225

southeast ends being higher in elevation and the north and south sides being lower inelevation. Piezometer 25 deep is on the south rim (Fig. 1) of the bay and has a lowerwater pressure than all piezometers north of it inside the bay. Piezometer 22 is on thenorth rim and has the lowest pressure of any other piezometer. Both of these piezometersare likely in zone 2. Piezometers 20 shallow and deep are located outside the bay to thenorth of the rim (north of Piezometer 22), in what looks like another Carolina Bay, andhave very low pressures starting in the late spring and lasting through the fall of 2002.Piezometer 11, in the middle of the bay, also has a low pressure but this is probably dueto the two main collector ditches located close by.

The top of the Black Creek Formation, a clay layer, is closest to the soil surface atcore 10, and then dips down towards the north and south sides of the bay. This is thesame way that the ground water in zone 2 is flowing. Data from piezometers in zone 3(Appendix A) suggest that ground water in this zone is moving in a southeast direction,towards the Atlantic Ocean. Piezometer 17B has a lower pressure than piezometer 1D,suggesting the above is true. Piezometer 17 may be in zone 2 or 3, it is unclear.

East of Juniper Bay is Shelley Bay. This may be drawing ground water in thezone 3 region towards it. Piezometer 1D, equipped with a pressure transducer, has alower pressure than piezometer 17. Piezometer 1B, west of piezometer 1D, has a lowerpressure than piezometer 1D. West of Juniper Bay is Hog Swamp, 30.48 meters abovesea level, and it may be drawing zone 3 ground water towards it. Juniper Bay is in themiddle of Hog Swamp and Shelley Bay, approximately 33.5 meters above sea level, andthe zone 3 ground water may be recharging them. Just west of Juniper Bay the elevationis about 38 meters above sea level, and core 19, located to the west of Juniper Bay (figure5), shows a sand layer approximately 4 meters thick from 28 to 32 meters above sealevel. This sand zone may be intercepted by Hog Swamp, approximately 30.5 metersabove sea level. The same situation occurs to the east of Juniper Bay with Shelley Bay.Cores 17, 26, and 28 all show mainly sand for meters down. Shelley Bay may bedrawing zone 2 and 3 ground water from this side of the bay. However, since there areonly three or four piezometers in zone 3, the exact direction of ground water flowing inthis zone cannot be precisely determined.

In Carolina Bays, reversal of the direction of ground water flow appearscommon, with net ground water outflow from the bay being the dominant direction (Lideet al, 1995). Where piezometer data show upward movement of water, ground waterdischarge is likely happening. Upward movement occurs in the northwest section of thebay, from piezometer 1S, located in zone 2, to well 1 (in zone 1) from mid winter throughthe summer of 2002 (Appendix A). This time period saw below average rainfallamounts (Table 1). Piezometer and well data show that the pressure is the highest in theshallow piezometer (4.5 m deep), a depth between those of the well (2.44 m) and thedeep piezometer (10.4 m). Ground water is likely entering the bay from the west side inzone 2 and can move upward and downward. Water moving downward runs into a thickclay layer that separates zone 2 from zone 3. A perched water table above this clay layermay exist.

226

Upward movement of ground water also occurs at the locations of piezometers 2Sand 2D, both of which are in zone 2, and well 2 (Appendix A). This nest is also in thenorthwest section of the bay, where ground water may be entering through zone 2.Upward movement was constant from the deep piezometer (4.4 m deep) to the shallowpiezometer (2.44 m deep) during 2002, indicating probable ground water inflow in thissection. Water pressure, and hence, direction between the well and the shallowpiezometer fluctuates from winter through mid-spring, 2002. From then on, theirpressures were approximately equal.

If the ground water and surface water have roughly the same pressure, then thethere is no real vertical movement between these two zones. Zone 1 is influencedhydrologically by rainfall while zone 2 may be influenced by direct rainfall and from rainfalling a certain distance from the bay and making its way to the zone 2 area and enteringthe bay laterally from outside. To the west of Juniper Bay is a water table divide.Rainfall falling to the east of the divide may make its way into Juniper Bay. Rainfallfalling on the west side of the divide will make its way towards Hog Swamp. Core 21,west of the bay and east of the water table divide, shows mainly sand from the surface to5.5 meters deep. Rainfall occurring in this area can move down through the sand intozone 2 and then flow into the bay.

During the winter and early spring of 2002, data from the piezometers indicatedthat the water pressure outside the bay to the north was equal to or above water pressureinside Juniper Bay (Appendix A). Manual measurements, however, showed that thepressure was always lower to the north of the bay. This discrepancy may be due tohuman error when pulling up the pressure transducers to download the data. The datafrom the manual measurements of piezometers 20S and 20D indicated that during the fallof 2001, the water level was the lowest of all measurements taken. This was during atime of below average rainfall (Table 1). During the winter months of 2002 the waterlevel increased, probably due to increased rainfall and a lack of ET. Piezometer 22, onthe north rim of the bay, had the lowest water pressure of all piezometers. Thepiezometer is about 7 meters deep. It is probably located in zone 2, but may interceptzone 3, leaking ground water from zone 2 into zone 3.

In Juniper Bay, from mid-spring through mid-fall of 2002, the water pressuredropped substantially more outside the bay to the north (where another Carolina Bay mayexist) than it did anywhere else inside or outside the bay. This may have been due to thatarea being heavily wooded and the rate of ET having been higher in the late spring andsummer, and rainfall having been less. Inside Juniper Bay there were no trees, except inthe south-eastern part where there is a stand of longleaf pine trees about 8 years old. Thewater table in that section, as shown in well 16A (Fig. 6), was much lower due to thedrawdown of the water table by those trees.

Juniper Bay may be a recharge area since it is not the lowest point on the locallandscape. Hog Swamp and Shelley Bay are both lower in elevation, as the Carolina Bayto the north, and ground water in Juniper Bay may be recharging them.

227

Ground water response to precipitation

Ground water levels in Juniper Bay began to respond immediately after a largerain event, but the rate of rise of the water levels is different between piezometers. Somepiezometers located in sandy material had water levels that rose quickly while some didnot, and some piezometers located in soil with more clay had water levels that rosequickly, while some did not.

One thing that may be influencing the ground water in the vicinity of piezometer2S is the main collector ditch. It may cut through the clay layer above this piezometerand into zone 2. If so, ground water in this region may flow towards it since it will be thearea lowest in pressure. An open ditch, if deep enough, attracts near-surface groundwater. This increase in ground water flow to the ditch increases the infiltration amountand rate of surface water (Dunn et al, 1996).

Hydraulic Conductivity

Data from the network of piezometers (Appendix A) in and around Juniper Baysuggest that ground water enters the bay primarily from the northwest side and thesoutheast side of the bay, both higher in elevation than the rest of the bay. Water flowstowards the lowest pressure head. The areas to the north and south of the bay are lowerin elevation. During the summer months of 2002, data suggest that ground water flowedout of the bay towards the north and possibly the south. Manual measurements of thewater level in the piezometers on the north-south transect show that the water level isconsistently lower outside the bay than inside the bay. Within the bay, ground waterflows towards the middle of the bay where piezometer 11 is located (Fig. 1). This maybe due to two larger ditches meeting near piezometer 11 and influencing the flow ofground water. Piezometer 11 is also in the lowest elevation in the bay.

Slug tests were performed in most piezometers to determine saturated hydraulicconductivity in order to calculate specific discharge, which is a velocity. Knowing thespecific discharge at specific locations will allow the approximate determination ofvertical and horizontal ground water flow patterns. Table 12 shows the saturatedhydraulic conductivity measurements obtained from analysis of the slug test data.

The slug tests show that the soil near piezometer 6 has the fastest hydraulicconductivity, meaning ground water flows faster at this location. The soil at this locationis sandy and water can move faster through the larger pores of the sand than through soilswith more clay and silt present in the texture.

228

Estimated Direction and Rates of Ground Water Flow

Table 13 shows the specific discharge between certain piezometers in zone 2.The first table shows the specific discharge in centimeters per hour. With these data theapproximate amount of ground water flowing into and out of the bay can be determined.

Table 13 shows that ground water flowed fastest from piezometer 3 to piezometer22. It is unclear, however, if the zones that the two piezometers are in are connected.The manual measurements also show a larger water level difference between piezometer3 and 22 than between 3 and 6. If the zones are connected then water is likely flowingfrom piezometer 3 to 22, meaning water is moving out of the bay at this location. Thedata also show that there may be some ground water movement from piezometer 3 to 6,since the water level, or pressure, is lower at 6 than at 3 (Appendix A). Flow becamefaster as the seasons changed from winter to spring to summer from piezometer 3 to 6, 6to 11, and 20 to 22. In many cases, however, flow decreased as the weather becamewarmer. This is probably due to the increase in ET and soil evaporation.

Ground water levels are higher inside the bay than outside the bay along thenorth-south transect (Appendix A). This feature of having higher ground water levelsinside the bay was also found in other Carolina Bay studies. This is due to the thick claylayer that exists below the surface of these bays. The clay layer below Juniper Bayshows up in the soil profiles between 6 and 10 meters deep. The outside areas to thenorth and south also represent where ground water may be moving out of the bay.

There are not as many piezometers along the east-west transect outside the bay asthere are along the north-south transect (Fig. 1). There is one to the west and one to theeast of the bay. There are also no soil cores from these two locations. This makes thedetermination of which zone they are in more difficult. The same pattern of lower waterlevels outside the bay in these two piezometers exists. If these two piezometers are inzone 3, then ground water in zone 2 can still be entering the bay from the west and eastand exiting to the north and south.

Hydrologic Restoration

The goal of the Juniper Bay mitigation project is to re-establish a stable wetlandsystem that will restore natural processes, structure, and species composition to mitigatefor wetland functions and values that will be impacted by highway construction activitiesin the Lumber River Basin (N.C.D.O.T., 2001).

Reference bay data are useful in the restoration process because they providetarget data on the hydrology of a natural Carolina Bay. As shown in Figure 7, thereference bays have the water table closer to the surface and for longer periods of timethan most wells in Juniper Bay, and the length of time that wetland hydrology occurs islonger than the 12.5% of the growing season criteria set by the Corps for wetland

229

restoration. This means that the target water table depth and duration for Juniper Baymay be longer than the restoration plan calls for.

230

REFERENCES

Bouwer, Herman. 1989. The Bouwer and Rice Slug Test – An Update. Ground Water,27(3), 304-309.

Bowen, Robert. 1986. Groundwater 2nd ed. Elsevier Applied Science Publishers,London, England.

Butler, James J. 1998. The Design, Performance, and Analysis of Slug Tests. LewisPublishers, Boca Raton.

Dunn, S.M., Mackay, R. 1996. Modeling the hydrological impacts of open ditchdrainage. Journal of Hydrology, 179, 37-66.

Fetter, C.W. 2001. Applied Hydrogeology, 4th ed. Prentice Hall, New Jersey.

Grant, John A., Brooks, Mark J., Taylor, Barbara E. New constraints on the evolution ofCarolina Bays from ground-penetrating radar. Geomorphology, 22, 325-345,1998.

Hillel, Daniel. 1998. Environmental Soil Physics. Academic Press, San Diego.

Horton, Jr., J. Wright, Zullo, Victor A. 1991. The Geology of the Carolinas: CarolinaGeological Society Fiftieth Anniversary Volume. The University of TennesseePress, Knoxville.

Hunt, Randy J., Krabbenhoft, David P. 1996. Groundwater inflow measurements inwetland systems. Water Resources Research, 32(3), 495-507.

Jarvis, N.J., Zavattaro, L., Rajkai, K., Reynolds, W.D., Olsen, P.-A., McGechan, M.,Mecke, M., Mohanty, B., Leeds-Harrison, P.B., Jacques, D. 2002. Indirectestimation of near-saturated hydraulic conductivity from readily available soilinformation. Geoderma, 108, 1-17.

Kao, C., Bouarfa, S., Zimmer, D. 2001. Steady state analysis of unsaturated flow abovea shallow water-table aquifer drained by ditches. Journal of Hydrology, 250, 122-133.

Mallants, Dirk, Mohanty, Binayak P., Vervoort, Andre, Feyen, Jan. 1997. Spatialanalysis of saturated hydraulic conductivity in a soil with macropores. SoilTechnololy, 10, 115-131.

Mitsch, William J., Gosselink, James G. 2000. Wetlands 3rd ed. John Wiley & Sons,Inc. New York.

231

North Carolina Department of Transportation (N.C.D.O.T.). 2001. Mitigation Plan forthe proposed Juniper Bay Wetland Mitigation Site. North Carolina Department ofTransportation Project Development and Environmental Analysis Branch NaturalSystems Unit, Raleigh, North Carolina.

Richardson, J.L., Vepraskas, M.J. 2001. Wetland Soils. Genesis, Hydrology,Landscapes, and Classification. Lewis Publishers, Boca Raton.

Schalles, John F., Shure, Donald J. 1989. Hydrology, community structure, andproductivity patterns of a dystrophic Carolina bay wetland. EcologicalMonographs, 59(4), 365-385.

Trettin, Carl C., O’Ney, Susan E., Eisenbies, Mark H., Miwa, Masato. 1999. Hydrologicprocesses in the vicinity of a Carolina Bay affecting water quality: An assessmentin associatin with a hardwood fiber firm. Southern Research Stationo, Center forForested Wetlands Research, Charleston, SC.

United States Global Change Research Information Office (G.C.R.I.O). 2002.http://www.gcrio.org/geo/karst.html.

Zanner, Bill, Farrell, Kathleen, Wysocki, Doug. 2001. Fitting Juniper Bay into theLandscape: Preliminary review based on what we have seen so far.

232

Table 1: Rainfall averages for Lumberton, N.C., Juniper Bay, and the Elizabethtown,N.C.Time Period Lumberton

averages, 1971-2000 (cm)

Juniper Bayaverages, 2001(cm)

Juniper Bayaverages, 2002(cm)

Elizabethtown,averages, 1971-2000 (cm)

January 10.9 - 9.1 11.1February 8.6 - 5.3 8.4March 10.9 9.4 7.7 11.5April 7.2 0.6 6.25 7.8May 10.1 6.2 3.8 9June 11.58 14.1 7.3 10.4July 14.24 17.8 9.7 14.7August 13.1 19.2 22 15.6September 11.7 6.2 3.4 13.6October 8.5 5.4 10.5 8November 6.8 4.6 8.4 6.8December 8.2 1.8 8.3 8.7Spring 28.9 20.9 17.35 27.2Summer 39.04 43.2 36.5 43.9Fall 23.5 11.8 27.3 23.5Winter 30.4 - 22.1 31Annual 121.9 85.4* 103.4 125.6* only includes data from March through December, 2001.

Table 2: Wells in Juniper Bay that meet the minimum wetland hydrology requirementsset by the Army Corps of Engineers.

Juniper BayWells that meet

minimum wetlandhydrology

requirements

Consecutive daysrequirements weremet, 2001 growing

season*

Consecutive daysrequirements weremet, 2002 growing

season*

Soil series;mineral, mineral

with histicepipedon, organic

Well 3 - 24 Leon, mineralWell 8 35 30 Leon, mineral with

histic epipedonWell 12 28 28 Leon, mineralWell 15 - 14 Ponzer Muck,

organicWell 16 28 28 Ponzer Muck,

organic* Growing season for project site is March 26 through October 30.

233

Table 3: Percentage of the soil at Juniper Bay that is mineral, mineral with a histicepipedon, and organic.

Juniper BayPercentage of Bay Soil Series Soil Type

25% Leon Mineral15% Leon Mineral with histic

epipedon10% Pantego Mineral20% Rutlege Mineral70% Total Mineral soil30% Ponzer Muck Organic

Table 4: Percentage of Juniper Bay that meets the U.S. Army Corps of Engineersrequirements for wetland hydrology.

Juniper BaySaturated to within 30 cm from soil surface for:

Percentage of: 5% of growingseason (11 days)

8.75% of growingseason (19 days)

12.5% of growingseason (27 days)

Juniper Bay 22% 15% 12%Mineral soil total 43% 43% 29%

Mineral soil: Leon 43% 43% 29%Mineral soil:

Pantego0% 0% 0%

Mineral soil: Rutlege 0% 0% 0%Mineral soil withhistic epipedon:

Leon

33% 33% 33%

Organic soil: PonzerMuck

29% 14% 14%

234

Table 5: Number of consecutive days the water table was within 30 cm from the soilsurface in the reference bay wells.

Wells that meet minimumwetland hydrology

requirements

Consecutive days during2002 growing season

(March 25 – November 4) thatthe water table is within 30

cm of soil surface*

Soil type (mineral,histic mineral,

shallow organic,deep organic)

Causeway Bay well 1** 4 mineralCauseway Bay well 2 1, 45***, 1, 4, 5, 13, 7, 1 histic mineralCauseway Bay well 3 3, 45, 8, 4, 4, 5, 4, 4 shallow organic

Charlie Long Bay well 1 3, 1 mineralCharlie Long Bay well 2 3, 45, 1, 8, 1, 21 histic mineralCharlie Long Bay well 3 3, 16, 5, 3, 2, 2, 2 shallow organicCharlie Long Bay well 4 3, 7 deep organic

Tatum Millpond Bay well 1 3, 21, 21, 25, 8, 21 mineralTatum Millpond Bay well 1A 3, 1, 4, 1, 2, 2, 1, 4, 1, 30, 21 mineralTatum Millpond Bay well 2 3, 21, 21, 18, 12, 1, 5, 21 histic mineralTatum Millpond Bay well 3 3, 21, 21, 18, 13, 5, 9, 21 shallow organicTatum Millpond Bay well 4 33, 21, 2, 19, 8, 3, 8, 21 deep organic* breaks in the wells meeting wetland hydrology requirements due to water tabledropping below 30 cm or lack of data for that period of time, as shown in graphs.

** wells in italics indicate there was no period that wetland hydrology requirementswere met.

*** bold numbers indicate an amount that meets the minimum wetland hydrologyrequirement of 5% set by the U.S. Army Corps of Engineers.

Table 6: Percentage of Charlie Long Millpond Bay in the three soil types.Charlie Long Millpond Bay

Percentage of Bay Soil Series Soil Type30% Lynn Haven/Leon Mineral20% Lynn Haven/Leon Mineral with histic

epipedon50% Total Mineral soil50% Pamlico Muck Organic

235

Table 7: Percentage of Causeway Bay in the three soil types.Causeway Bay

Percentage of Bay Soil Series Soil Type7% Lynn Haven/Leon/Kureb Mineral3% Lynn Haven/Leon/Kureb Mineral with histic

epipedon10% Total Mineral soil90% Pamlico Muck Organic

Table 8: Percentage of Tatum Millpond Bay in the three soil types.Tatum Millpond Bay

Percentage of Bay Soil Series Soil Type20% Lynn Haven/Torhunta Mineral10% Lynn Haven/Torhunta Mineral with histic

epipedon30% Total Mineral soil70% Pamlico Muck/Croatan

MuckOrganic

Table 9: Percentage of the soil types in Causeway Bay that meet the wetland hydrologyrequirements in objective one.Causeway Bay

Saturated to within 30 cm from the soil surface for:Percentage5% of growing

season (11 days)8.75% of growingseason (19 days)

12.5% of growingseason (27 days)

Whole Bay 93% 93% 90%Mineral soil 0% 0% 0%

Mineral soil withhistic epipedon

100% 100% 0%

Organic soil 100% 100% 100%

236

Table 10: Percentage of the soil types in Charlie Long Millpond Bay that meet thewetland hydrology requirements in objective one.Charlie Long Millpond Bay

Saturated to within 30 cm from the soil surface for:Percentage5% of growing

season (11 days)8.75% of growingseason (19 days)

12.5% of growingseason (27 days)

Whole Bay 70%? 20%? 20%?Mineral soil 0%? 0%? 0%?

Mineral soil withhistic epipedon

100% 100% 0%

Organic soil 100% 0% 0%

Table 11: Percentage of the soil types in Tatum Millpond Bay that meet the wetlandhydrology requirements in objective one.Tatum Millpond Bay

Saturated to within 30 cm from the soil surface for:Percentage5% of growing

season (11 days)8.75% of growingseason (19 days)

12.5% of growingseason (27 days)

Whole Bay 100% 100% 100%Mineral soil 100% 100% 100%

Mineral soil withhistic epipedon

100% 100% 100%

Organic soil 100% 100% 100%

Table 12: Saturated hydraulic conductivity values obtained through slug tests.Piezometer Saturated hydraulic conductivity

1-S 0.19012-S 0.03672-D 0.19973 0.05826 0.3884

10-S 0.018510-D 0.201311 0.026512 0.051217 0.0758

20-S 0.087120-D 0.264222 0.0374

25-S 0.151525-D 0.0435

237

Table 13: Specific discharge calculated from saturated hydraulic conductivity values.

Specific discharge, v, cm/hr from: Fall 2001 Winter 2002 Spring 2002 Summer 2002 P2 to P1 0.000057 0.000053 0.000057 P3 to P6 0.000108 0.000124 0.000162 0.000143 P10 to P6 0.000146 0.000137 0.000148 0.000077 P6 to P11 0.000091 0.000117 0.000155 P10 to P11 0.000094 0.000117 0.000081 P12 to P11 0.000042 0.000042 0.000031 P20 to P22 0.000508 0.000512 0.000560 P10 to P25 0.000273 0.000266 0.000263 P2 to P6 0.000236 0.000207 0.000120 P3 to P22 0.000766 0.000794 0.000891

238

12

2A

5

5A

3

6

7

810

15

11

16

12

17

25

20

22

16A

25A25B

1B 2B

8A

8B

17B16B

23A

Well Locations

Piezometer Locations

Figure 1. Map of piezometer and well locations.

239

Figure 2. Soil core locations.

240

Bottom Clay layer

Perimeter ditch Main outflow ditchLateral ditch

“Middle” clay layer

2.44 - 3.05 m(8 - 10 ft)

3.05 - 3.66 m(10 - 12 ft)Zone 1

Zone 2

Zone 3

Figure 3. Internal system boundaries. Clay layers break sediments into threedistinct zones or aquifers.. Each zone consists largely of sandy deposits and isresponsible for most ground water movement. Ditches currently penetrate into zone1. A few deep ditches may penetrate into zone 2. Most lateral moving ground wateris believed to move through zone 2 at this time. Zone 3 is believed to be uninvolvedin the current hydrology of Juniper Bay.

241

Figure 4: Wells located in the northwest section of Juniper Bay

-240

-210

-180

-150

-120

-90

-60

-30

0

30

90 130 170 210 250 290 330 370 410 450 490 530 570 610 650 690 730

Day of Year from April, 2001

Dep

th t

o th

e w

ater

tab

le (

cm)

Well 1

Well 1A

Well 2

Well 2A

Well 5

Well 5A

Well 23A

242

Figure 5: Wells located in the middle of Juniper Bay

-240

-210

-180

-150

-120

-90

-60

-30

0

30

90 130 170 210 250 290 330 370 410 450 490 530 570 610 650 690 730

Day of Year from April, 2001

Dep

th t

o th

e w

ater

tab

le (

cm) Well 3

Well 6

Well 7

Well 10

Well 11

Well 15

Well 20

Well 25

Well 25A

Well 25B

243

Figure 6: Wells in the Southeast section of Juniper Bay

-240

-210

-180

-150

-120

-90

-60

-30

0

30

90 130 170 210 250 290 330 370 410 450 490 530 570 610 650 690 730

Day of Year from April, 2001

Dep

th t

o th

e w

ater

tab

le (

cm)

Well 8

Well 8A

Well 8B

Well 12

Well 16

Well 16A

Well 16B

Well 17

244

Figure 7: Wells in the reference bays (CB=Causeway Bay, CL=Charlie Long Millpond Bay, TM=Tatum Millpond Bay)

-125

-100

-75

-50

-25

0

25

440 460 480 500 520 540 560 580 600 620 640 660 680

Day of Year from April, 2001

Dep

th t

o th

e w

ater

tab

le (

cm)

CB 1

CB 2

CB 3

CB 4

CL 1

CL 2

CL 3

CL 4

TM 1

TM 1A

TM 2

TM 3

TM 4

245

Appendix A: Piezometer data.

Nest 1 hydrograph December 20, 2001 to November 13, 2002

-210

-180

-150

-120

-90

-60

-30

0

30

350 370 390 410 430 450 470 490 510 530 550 570 590 610 630 650 670 690

Julian day

wat

er d

epth

(cm

)

P-1-D 34.1'

P-1-S 14.8'

Well 1 8'

246

Nest 2 hydrograph December 20, 2001 to November 13, 2002

-140

-120

-100

-80

-60

-40

-20

0

20

350 370 390 410 430 450 470 490 510 530 550 570 590 610 630 650 670 690

Julian Day

Dep

th o

f w

ater

(cm

)

P 2-D 14.5'

P 2-S 8.1'

Well 2 8'

247

Piezs in west to east transect corrected for elevation

-300

-290

-280

-270

-260

-250

-240

-230

-220

-210

-200

-190

-180

-170

-160

-150

-140

-130

-120

340 350 360 370 380 390 400 410 420 430 440 450 460 470 480 490 500 510 520 530 540 550 560 570 580

Julian Day

Dep

th t

o w

ater

(cm

)

P-1-S 16.41'

P-11 16.78'

P-12 17.73'

P-2-D 16.03'

P-6 21.48'

2-S cor

248

Piezometers in north to south transect in Aquifer 2 region

-340

-330

-320

-310

-300

-290

-280

-270

-260

-250

-240

-230

-220

-210

-200

-190

-180

-170

-160

-150

340 350 360 370 380 390 400 410 420 430 440 450 460 470 480 490 500 510 520 530 540 550 560 570 580

Julian Day

Dep

th t

o w

ater

(cm

)

10-D cor

3 cor

6 cor

20-S cor

20-D cor

249

P-22 from December 20, 2001 to November 13, 2002

-400

-350

-300

-250

-200

-150

-100

-50

0

340 360 380 400 420 440 460 480 500 520 540 560 580 600 620 640 660 680 700

Day of year

Dep

th t

o w

ater

(cm

)

250

Chapter 11

WATER BUDGET FOR JUNIPER BAY

G.S. Kreiser, R.L. Huffman, and M.J. Vepraskas

INTRODUCTION

Hydrology is the most important variable in the creation and maintenance ofdifferent types of wetlands and wetland processes (Mitsch and Gosselink 1993). For theclassification, assessment, and restoration of wetlands, there is a need to know theamounts and timing as well as the sources of water (Owen 1995). Even thoughhydrology is known to be important, it is often overlooked and the least understoodaspect of wetlands. This may be due to the fact that assessing hydrology is a complexand time-consuming process.

Early wetland studies concerning hydrology dealt with the relationship betweenwetland productivity and species composition (Mitsch and Gosselink 1993). In the lasttwenty years, numerous papers have dealt with the various broad aspects of hydrology inwetlands (Carter et. al 1978; Carter 1986; Carter and Novitzki 1988). However, therehave been few studies that have described the detailed hydrologic characteristics withinspecific types of wetlands. Exceptions to this include studies of northern peatlands, saltmarshes, cypress swamps, and large-scale wetland complexes (Mitsch and Gosselink1993).

Hydrologic studies must be well planned in order to quantify the temporal andspatial distribution of water, and must consider all possible inputs and outputs of awetland. This process for accounting all of the water sources and sinks within a definedsite is commonly called a water budget (Roig 2000). Figure 1 illustrates the componentsof a wetland. Water budget equations are often used in detailed hydrologic assessment ofwetlands (Rykiel 1984; Hyatt and Brooks1984). Water budgets are also useful for thecalculation of nutrient budgets. In addition, they can be used to estimate unknownhydrologic components such as groundwater flow and for the prediction of the effects ofnatural and anthropogenic changes on water inputs and outputs (Carter 1986; Roig 2000).

The general components of a water budget equation showing the water storage,inflows, and outflows of a wetland may be expressed as (Mitsch and Gosselink1993):

?V/?t=Pn+Si+Gi-ET-So-Go±T [1]

where:?V/?t = change in volume of water storage in wetland per unit time, tPn = net precipitationSi = surface inflows, including flooding streamsGi = groundwater inflowsET = evapotranspiration

251

So = surface outflowsGo = groundwater outflowsT = tidal inflow or outflow

It is important to note that not all variables occur in all wetlands. There are manydifferent forms of this equation, all of which are essentially the same (Carter et. al. 1978;Roig 2000).

The water budget (Eq. 1) at first glance seems deceptively simple, where waterinputs equals water outputs, plus or minus the change in storage, but care must be takenwhen using it to establish the hydrology of a wetland (Winter 1981). The central problemin determining water budgets for a wetland is how well the individual inputs, outputs, andchange of storage can be measured or estimated and the magnitude of the associatederrors (Dooge 1972; Winter 1981; Carter 1986). For most hydrologic studies, it isdesirable to measure or estimate all of the components in order to calculate a waterbudget (Dooge 1972; Hyatt and Brook 1984; Carter 1986). However, this is not alwayspossible due to the difficulties in making hydrologic measurements, and one componentis calculated as the residual of the water budget equation (Eq. 1).

The inherent problem with the residual component is that it contains the sum of allerrors from the other terms in the budget. Errors can be classified into two categories ofeither measurement or interpretation (Winter 1981). Measurement errors occur fromtrying to take measurements using imperfect instruments, inadequate sampling design,and data collection procedures. Interpretation errors occur as a result from using pointdata in order to estimate quantities for a longer period of time (Winter 1981). Theseerrors can have a significant effect on the calculations of a water budget. However, erroranalysis is not commonly used and the residual term is given a great deal of interpretationand importance, even though the residual term has little meaning. Winter (1981)recommends that any hydrologic budget, however derived, include error analysis in orderto allow for realistic use of water budgets. By including error analysis Equation 1becomes:

?V/?t=Inputs-Outputs±error [2]

The inputs and outputs are the same as in Equation 1. Error is calculated from thestandard deviations of measurement and the known instrument error and then is summedup in the final water budget equation (Owen 1995).

Water budgets are useful because they provide “a first approximation of inputs andoutputs as a basis for nutrient balance and energy studies, hydrologic models, andpredictions of impact” (Carter 1986). Carter (1986) notes that water budgets arecommonly used in wetland analysis. While there are known errors associated withcalculating a water budget, water budgets provide an initial base for further detailedanalyses of wetland hydrology.

252

In order to determine the sources of water occurring at Juniper Bay, the hydrologyis being studied. To determine the hydrologic inputs and outputs into the bay, a waterbudget is being used. The objectives of this research are to establish a water budget,compute the magnitude of water inflow and outflow into the bay, and predict possibleimpacts of the restoration.

MATERIALS AND METHODS

The North Carolina Department of Transportation (NCDOT) is in the process of restoring adrained Carolina Bay in Robeson County. The project will provide compensatory wetlandmitigation in the Lumber River Basin of southeastern North Carolina, which will offset wetlandimpacts from road construction projects in the river basin (Hauser 2001). The site, known asJuniper Bay (Figure 2), is composed of 300 hectares of an extensively drained Carolina Bay thatwas used for agricultural production. The goal of the restoration project is to restore the siteback to a wetland, thus recreating the functions and values of a Carolina Bay. As part of thisrestoration project, North Carolina State University is investigating the hydrologic, soil andvegetative changes that occur in Juniper Bay as a result of this restoration project.

Juniper Bay is located approximately 16 kilometers southeast of Lumberton, NC(Figure 3). The bay contains an extensive network of ditches that were installed in orderto drain the bay for agricultural production (Figure 4). All of the surface water exits thebay through a perimeter ditch and a large main ditch. The soils of the bay were classifiedduring the summer of 2001 with Juniper Bay having both organic and mineral soils.

Precipitation

Precipitation was measured using three tipping bucket gauges that were locatedwithin the bay (Fig. 4). Gauges located in the NW and SE measured rainfall to thenearest .01- inch, while the gauge at the weather station measured rainfall amounts to thenearest 0.1-millimeter. The rain gauges at NW and SE were Davis Instruments, RainCollector II. The event recorder used a HOBO Event Logger, by ONSET ComputerCorporation. The rain gauge at the weather station was Texas Electronics, Inc.TE525MM. Monthly totals for all rain gauges were computed and then compared.Gauges that obviously had inaccurate totals for a month were discarded. The remaininggauges were then used to determine the average monthly total of precipitation for theentire bay.

Evapotranspiration

Data that were used to estimate PET were collected at the weather station locatednear the center of bay (Figure 4). Measurements included hourly readings of directradiation, net radiation, wind speed, and wind direction. Wind speed and wind directionwere measured with Climatronics Wind Speed and Direction Sensors (Model CS 800-L).Net radiation was measured with a REBS Net Radiometer (Model Q7_1). Directradiation was measured with a LI-COR LI200SZ Silicon Pyranometer. Air temperatureand relative humidity were measured using Valsala-Temperature and RH probe ModelHMP45C. All measurements recorded were stored in a Campbell Scientific CR-10X datalogger.

253

Reference ET was calculated using software from University of Idaho at Kimberly(University of Idaho at Kimberly 2002). The reference crop was assumed to be a shortcool-season grass, which gave a reference ET (ETo). Potential evapotranspiration wascalculated by using the Penman-Monteith equation where:

ET = ?(Rn-G) + ?acp(eoz-ez)/ra [3]

? +?(1+ rs/ra)where:Rn = net radiationG = soil heat flux? = slope of satruation vapor pressure? = psychrometric constanteo

z = mean saturation vapor pressurers = bulk surface resistancera = aerodynamic resistanceez = actual vapor pressure of air?a = air densitycp = specific heat of dry air

The hourly weather data was used to calculate hourly ETo values, which were thensummed to get a monthly total. A crop cover coefficient was applied in order to getestimated ETc (using equation ETc=EToKc). A crop (cover) coefficient was constructedusing the procedure outlined in ASCE Hydrology Handbook (1996). The Kc curve wasconstructed by dividing the growing season into four parts that describe the growthstages: the initial period, development period, midseason period, and late season period.

One of the dominant plant species at Juniper Bay during the study was Dog Fennel(Eupatorium capillifolium L.). Dog Fennel is an aggressive weed species that is native tothe southeastern United States, which commonly occurs on sites such as old fields,ditches, and disturbed pastures (Van Deelen 1991). Dog Fennel is often described as anannual that can grow to heights of 4 to 5 feet (1.2-1.5 m). Using information from theASCE Hydrology Handbook (1996), a crop similar to dog Fennel was picked in order toget the crop coefficient. Millet (Panicum miliaceum L.) was listed as a crop with amaximum height of 1.5 m, which is similar to dog fennel. Using the numbers for millet,a Kc curve (Figure 5) was developed for Juniper Bay. Using the Kc graph, cropcoefficients were picked for each month, which are listed in Table 1. Using the referenceequation of ETo multiplied by Kc will provide a monthly estimated ETc for Juniper Bay.

Surface Outflow

Surface outflow was measured at the main outlet using dual compound weirs(Figure 4). Dual compound weirs were selected because they measure both low and highflow events, and also because a cement structure was already in place to support twoweirs. Compound weirs consist of a rectangular notch with a V-notch cut into the centerof the crest. Determining discharge from compound weirs requires the use of twodifferent equations. Which equation is used depends on whether the discharge iscontained in the V-notch or rectangular portion of the weir. If there are high flow events

254

that are contained in the rectangular portion of the weir then the following equation isused (USBR 2003):

Q=3.9h11.72-1.5+3.3Lh2

1.5 [4]Where:Q= discharge in ft3/sh1= head above the point of the V-notch in ftL = combined length of the horizontal portions of the weir in fth2= head above the horizontal crest in ft

However, if the flow is confined to only in the V-notch portion of the weir then thestandard V-notch equation is used (USBR 2003)

Q = 2.49 x h12.48 [5]

In order to determine the stage of the water, pressure transducers were locatedabove and below the weir. Knowing the elevation of the weir invert and the height of thewater above the weir gives the height of water that is passing over the weir. This heightof water determines what part of the weir the water is in and this then determines whichequation is used. A Campbell Scientific CR-10X is used to run the calculations and tostore the data. Surface outflow volumes were converted to equivalent depths of water bydividing by the surface area of the bay. The outflow measurements were then summed inorder to get monthly totals.

During failure of the primary outflow measurement system, surface outflow wasestimated by comparing the water elevations during the failure with water elevationsduring non-failure periods. This assumes that similar water elevations would havesimilar outflow amounts. Using this assumption, surface outflow was estimated whenthere was system failure.

Water Table Depth

Sixteen automatic monitoring wells (Remote Data Systems, Inc., Wilmington, NC)were installed inside the bay (Figure 6). Wells were installed by boring a hole 4 inches indiameter to a depth of 80 inches. The area around the well screen was back filled withsand in order to prevent clogging of the screen. A layer of bentonite was placed on top ofthe sand layer in order to seal the well from surface water. A conical mound of soil wasthen placed on top of the bentonite as a preventive measure to keep surface water awayfrom the well. Water table depths each well were recorded hourly. In order to getaverage water table depth for the whole bay, all the well depths were averaged together.The average water table depth was then used in calculating the change in storage of thebay for each month.

Change in Storage

Change of storage was determined from the differences in water table depth at thebeginning and end of each month. Change in storage was determined by multiplying the

255

change in the average water table depth by the average drainable porosity of the soils.Drainable porosity is the amount of water drained for a given drop in the water table(University of Minnesota 2003). Drainable porosity is affected by the soil texture andstructure, with sands having larger drainable porosities and clays having smallerdrainable porosities (University of Minnesota 2003). Most of the soils in Juniper Bayhave been classified as either sandy clay or sandy loam (Ewing and Vepraskasunpublished). According to the FAO Irrigation and Drainage Paper 38 (1980) thedrainable porosity for these types of soils is between 0.03 and 0.12 cm3 /cm3. A drainableporosity of 0.1 was used as an average drainable porosity of the whole bay. Multiplyingthe average drainable porosity by the change in water table depth for each month gave themonthly change in storage for Juniper Bay.

Groundwater

Groundwater inflow and outflow were estimated as the residual of the water budgetequation (Eq 1). Using a modified version of Equation 1 groundwater was calculatedwhere:

Inputs-Outputs=?Storage?Storage=(Precip + Gi) –(ET+Go+So)Precip + Gi = ET + Go +So + ?StorageGi-Go = ET +So + ?Storage – Precip

The groundwater component of the water budget is the net groundwater movement in thebay and it was estimated for each month. The term Gi-Go was computed as a lengthmeasurement, but can be converted to a volume by multiplying by the area of the bay.

Statistics

Statistical software (SAS) was used to perform summary statistics and multivariatecorrelation on all components of the water budget.

Results and Discussion

Precipitation

Average monthly precipitation amounts are shown in Figure 7. These are theaverage values determined from the rain gauges at Juniper Bay. Long-term normalrainfall levels are also shown in Figure 7. Seven months had lower than normal rainfall,particularly from April through July. Rainfall after August was generally normal orwetter than normal.

Evapotranspiration

There was a seasonal trend in the amount of PET (Figure 8). The total amount ofET was estimated at 845 mm for the year. This seems to be a reasonable estimate sinceothers have estimated ET in the coastal area of North Carolina to be approximately 900

256

mm (van Bavel and Verlinden 1956). Evapotranspiration has a negative correlation withsurface outflow (R2= -0.5805). In months when there was high ET there was a lowamount of surface outflow. In the summer months of June, July and August the highestrates of ET were calculated for Juniper Bay. During these months there werecorresponding low amounts of surface outflow. ET is the biggest sink and accounts for54 percent of all water leaving the bay.

Change in Storage

Throughout the year the monthly change in storage was quite variable as expected(Table 2). A negative number represents a drop in the average water table at Juniper Bayduring that month and a positive number indicates a rise in the water table (Figure 9). Thetwo main driving forces for the fluctuation of the change in storage appear to beprecipitation and ET. Table 2 shows that when ET was greater than precipitation duringthe month, in most instances, there was a negative change in storage. Likewise, whenthere was more precipitation than ET during the month, the change in storage waspositive, representing a rise in the water table. However, statistical analysis did not showthis relationship to be significant. There was no significant correlation between change instorage and ET (R2= -0.0127). There was however, a significant correlation between thechange in storage and the amount of rainfall during a particular month (R2=0.9364). Thisseems reasonable, as precipitation is the main hydrologic input into Juniper Bay. Sinceprecipitation is the main input into the system it is logical that the system would respondthe most to precipitation. A possible explanation as to why ET was not statisticallysignificant may be that precipitation is the main driving force for change of storage andonly when there is little rainfall will ET cause a drop in the water table. The smallamount for change in storage is consistent with the fact the over a long time period thechange in storage equals about zero; that is, there is almost no net change in the watertables.

Surface Outflow

Monthly surface outflow was also quite variable and followed a trend opposite tothat of ET (Figure 10). There was failure of the primary system during a two-week periodin April of 2002 that required the estimation of outflow during that period. Outflow wasestimated and added to the monthly total for April. The summer months had the lowestoutflow and amounts increased in the fall and winter. During the summer months therewas little surface outflow when there was little precipitation and high amounts of ET.There were higher outflow amounts during the fall and winter when there was adequaterainfall and lower ET rates.

Net Groundwater Input

The groundwater input and output were estimated as the residual of the rest of thewater budget equation as mentioned in the method section. For this reason, groundwateris calculated as the net groundwater input, which could be either positive or negative.

257

The net groundwater input was calculated as 563 mm for this year. This means thatmore groundwater was entering the bay than leaving and the excess groundwater wasleaving as surface outflow. When the groundwater component was negative, moregroundwater was leaving the bay than entering, and Juniper Bay was acting as a rechargewetland. Figure 11 shows that for most of the year Juniper Bay had a positive netgroundwater movement and was acting as a discharge wetland. The net groundwatercomponent accounts for 35 percent of the total water inputs into Juniper Bay. Thegroundwater component is large suggesting that Juniper Bay has a significantgroundwater input.

Error Analysis

All components in the water budget that were measured contain errors that arepropagated through the water budget equation and are contained within the residual term;in this case groundwater. Significant errors in interpretation can occur if the residualterm is used without respect to errors that are inherent to it. Table 3 lists all theassociated errors for each component and the range. Errors were taken from the literature(Winter 1981; Owen 1995). All the errors combined equal to 171 mm. If all the errorsare either added or subtracted, the range for groundwater is 392 mm to 734 mm. Evenwith all the errors associated with using the residual, the groundwater component is stillpositive indicating net groundwater movement into Juniper Bay.

The ET calculation has by far has the largest potential for error, and could have thebiggest effect on the estimate of groundwater input. Taking that into consideration, ETwas calculated with different percent error and graphed with the assumption that all othercomponents remained the same. Figure 12 shows that if ET calculations were off by fiftypercent there would still be a net positive groundwater component of 141 mm for theyear.

Groundwater Flow into Juniper Bay

After taking into consideration all of the errors associated with the water budget,even in the worst-case situation of being off by fifty percent in the ET calculation therestill is a positive net groundwater component. This raises the question as to where thegroundwater comes from. In order to determine the possible sources of groundwater it isimportant to look at the topography of Juniper Bay and the surrounding area.

Juniper Bay lies at an intermediate landscape position, at 112 feet above sea level(Figure 13). Uplands occur to the east and west of Juniper Bay with elevations of about124 to 131feet. Lowlands occur to the south and north of Juniper Bay with elevationsaround 112 feet. These differences in elevation create a hydraulic gradient that mightaccount for groundwater inputs into the bay. The higher elevations or uplands would bepossible sources for groundwater inputs and the lower elevations would be possiblelocations for groundwater output.

258

Two possible sources for groundwater into Juniper Bay are local and regionalgroundwater. The difference between precipitation (1014 mm) and ET (845 mm)provides an estimate for the amount of water recharging the groundwater. In this case,169 mm of water has the potential to become groundwater for this year. The volume ofwater entering Juniper Bay was figured as the amount of groundwater multiplied by thebay’s area. The recharge area needed to supply this amount of water was calculated asthe volume of the bay divided by the potential amount of groundwater.

The adjacent uplands are the areas to the east and west of Juniper Bay and onlyaccount for about 16 percent of the area needed to meet the water requirements forgroundwater. The only other possible source for groundwater is from more distant orregional sources. If that is the case then the regional sources account for over 80 percentof the groundwater entering Juniper Bay.

If actual ET were 50 percent of ET then the net groundwater input would be less. Ifthe PET calculations were equal to ET then there would be a groundwater inflow of 563mm a year, which accounts for 35 percent of the total inputs into the bay (Figure 12).However, if the ET calculations were off by 50 percent then there would be groundwaterinflow of 141 mm a year, which accounts for about 11 percent of the total inputs intoJuniper Bay. Even if there is a 50 percent error in the ET calculations there still is agroundwater inflow that could be significant in the restoration project.

Monitoring wells located in the east and west of Juniper Bay suggest that there isgroundwater inflow into the bay. Wells 1 and 17 are in locations where it is thought thatthere is groundwater inflow. During a two-week period of no rainfall in August of 2002the wells show evidence of groundwater inflow (Figure 14 and 15). These two figuresshow that during the day there is a draw down in the water table that is caused by ET. Atnight, when there is no ET, the water level rises due to groundwater input from outsidethe bay. For a well that is located further inside Juniper Bay there is not that cyclicpattern of water table falling and rising in twenty-four hours. Figure 16 shows well 10that is located near the center of the bay, and unlike the other two wells there is not areplenishing of the groundwater. For well 10 during the same time period there is asteady drop in the water table caused by ET.

Other evidence of groundwater inflow into Juniper Bay is that there are organicsoils located in the center of the bay (Figure 17). Organic soils form when soils aresaturated long enough at the surface for significant organic matter to accumulate. Thissuggests that where there are organic soils in Juniper Bay, breaks in the clay layer allowfor the upwelling of groundwater. Early findings of ground penetrating radar have shownthat these areas of organic soils in Juniper Bay do have an indication of breaks in the claylayer (Szuch, personal communication 2003).

Implications

The water budget analysis suggests that Juniper Bay is currently acting as adischarge wetland, where groundwater is entering the bay and leaving as surface outflow.

259

Topographically Juniper Bay lies at an intermediate elevation and in a natural setting itcould be acting as a flowthrough wetland. Figure 18 represents the topographicrelationship of Juniper Bay to the surrounding areas.

There are two possible scenarios after restoration for what effect groundwater willhave on restoration. If the perimeter ditch is left open, then wetland hydrology can berestored, and the excess water will be taken away as surface outflow. The bay willcontinue to function as a discharge wetland. Figure 19 shows that with the main ditchclosed and the perimeter ditched left open, the perimeter ditch can still remove the excessgroundwater.

However, if the perimeter ditch is closed (Figure 20), then there will be no outlet forthe groundwater, and the water levels might rise above the present soil surface. Waterlevels may rise because it has been estimated that the subsidence of organic soils could beas great as 80 cm (Ewing and Vepraskas 2003). Which means that previously the soilsurface at Juniper Bay was higher. Restoration of Juniper Bay with the perimeter ditchclosed could possibly raise the water table above the surface and also might causegroundwater outflow, which might raise the water table in the surrounding area. Suchinstances of hydrological trespass could be a potential problem for the restoration project.

CONCLUSIONS

This study utilized a water budget in order to determine the potential of a drainedCarolina Bay to be restored into a wetland. Measured components of the water budgetincluded: precipitation, ET, surface outflow, and change in storage. Groundwater wasestimated as the residual from the water budget equation.

A water budget for Juniper Bay was estimated for 2002 to 2003. There was asignificant relationship between rainfall and change in storage as well as a negativecorrelation between surface outflow and ET. In addition, there was a positiverelationship between surface outflow and groundwater. These findings indicate thatJuniper Bay has a significant groundwater component and that this groundwater inflowcould possibly influence the restoration project.

Two possible scenarios upon restoration of Juniper Bay must be evaluated in lightof this information. In scenario one, the main ditch is plugged while the perimeter ditchwould remain open. Under this scenario, wetland restoration would occur, and theperimeter ditch would intercept the excess groundwater.

Under scenario two, the perimeter ditch would be closed resulting in groundwaterentering the bay and possibly raising the water table above the soil surface. This excesswater might cause groundwater outflow into the lower surrounding areas. The movementof groundwater into neighboring areas could possibly cause hydrologic trespass.

Upon restoration, it appears that Juniper Bay will function as a flowthrough wetlandwith the groundwater entering and leaving the bay. These findings indicate the

260

restoration project must take into account the groundwater component of the site. If themitigation project does not address the ground water component, it may have adverseeffects to the surrounding areas even if it provides mitigation credits for NCDOT.

261

REFERENCES

America Society of Civil Engineers. 1996 ASCE Manuals and Reports on Engineering Practice No. 28. Hydrology Handbook 2nd edition.

Carter, V. 1986. An Overview of the hydrologic concerns related to wetlands in theUnited States. Can. J. Bot. 64:364-374.

Carter, V., M.S. Bedinger, R. P. Novitzki, and W. O. Wilen. 1978. Water resources andwetlands In Wetland Functions and Values: The State of Our Understanding, P.E.Greeson, J.R. Clark, and J.E. Clark eds., American Water Resources Assoc., Minneapolis, Minn., p. 344-376.

Carter, V., and R. P. Novitzki. 1988. Some comments on the relation between groundwater and wetlands, In The Ecology and Management of Wetlands, D.D. Hook etal., eds., vol. 1: Ecology of Wetlands, Timber Press, Portland, Oregon, p. 68-86.

Dooge, J. 1972. The water balance of bogs and fens (review report). In Hydrology ofmarsh-ridden areas. Proceedings of Minsk Symposium, 1972. UNESCO Press, Paris. p. 233-271.

Ewing, J., and M.J. Vepraskas. 2003. Secondary Subsidence in Organic Soils.Unpublished.

FAO. 1980. Irrigation and Drainage Paper 38. Food and Agriculture Organization ofthe Unite Nations, Rome.

Hauser, J. NCDOT Develops Juniper Bay Mitigation Site. Centerline An EnvironmentalNews Quarterly, From the NCDOT Natural Systems Unit. January 2001 IssueNo. 4.

Hyatt, R. A., and G. A. Brook. 1984. Ground water flow in the Okefenokee Swamp andthe hydrologic and nutrient budgets for the period August, 1981 through July,1982. In The Okefenokee Swamp. A.D. Cohen, D.J. Casagrande, M.J. Andrejko,and G.R. Best, eds. Wetland Surveys, Los Alamos, NM. p 229-245

Mitsch, W.J., and J.G. Gosselink. 1993. Wetlands. 2nd edition. John Wiley & Sons,New York. p 67-111.

Owen, C.R. 1995. Water budget and flow patterns in an urban wetland. Journal ofHydrology 169:171-187.

Roig, L.C. 2000. Determining Existing Hydrologic Conditions, In WetlandsEngineering Handbook, US Army Corps of Engineers. p. 2-46.

262

Rykiel, E. J. Jr. 1984. General hydrology and mineral budgets for Okefenokee Swamp.In The Okefenokee Swamp. A.D. Cohen, D.J. Casagrande, M.J. Andrejko, andG.R. Best, eds. Wetland Surveys, Los Alamos, NM. p 212-228.

Sharitz, R.R., and J. W. Gibbons. 1982. The Ecology of Southeastern Shrub bogs(Pocosins) and Carolina Bays: A Community Profile.

Szuch, R. Personal communication, June 2003.

University of Idaho Kimberly Research and Extension Center. 2002. REF-ET ReferenceEvapotranspiration Software. Available at: http://www.kimberly.uidaho.edu/ref-et.

University of Minnesota Extension Service. 2003. Agricultural Drainage-PublicationSeries Water Concepts. Available at:http://www.extension.umn.edu/distribution/cropsystems/DC7644.html.

United States Bureau of Reclamation (USBR) Water Measurement Manual. Chapter 7Weirs. Available at http://www.usbr.gov/wrrl/fmt/wmm/chap07.html.

van Bavel, C.H.M., and F. J. Verlinden. 1956. Agricultural drought in North Carolina:North Carolina Agricultural Experiment Station Technical Bulletin 122, 60 p.

Van Deelen, Timothy R. 1991. Eupatorium capillifolium. In: U.S. Department ofAgriculture, Forest Service, Rocky Mountain Research Station, Fire

Sciences Laboratory (2003, June). Fire Effects Information System, [Online].Available: http://www.fs.fed.us/database/feis/eupcap/.

Winter, T. C. 1981. Uncertainties in estimating the water balance of lakes. Water ResourBull. 17:82-115.

263

Table 1. Crop Coefficient for each month taken from the Kc curve inFigure 6.The crop coefficient term is used in the reference ET equation ofETc=KcETo

Month Kc

February 0.3March 0.3April 0.4May 0.6June 1July 1

August 1September 0.8

October 0.4November 0.3December 0.3

January 0.3February 0.3

264

Change inStorage

PotentialEvapotranspiration

SurfaceOutflow

PrecipitationNet

Groundwater

mm mm mm mm mmFebruary -9.52 23.10 32.00 52.96 -7.38

March 12.89 32.70 60.00 106.17 -0.58April -60.38 58.80 64.50 39.54 23.38May -12.25 93.00 22.20 43.05 59.90June -9.42 164.00 3.40 70.10 87.88July -12.17 169.00 0.06 59.94 96.95

August 87.26 133.00 4.90 171.07 54.09September -55.19 66.64 19.30 33.91 -3.16

October 43.60 29.86 97.38 147.20 23.64November 3.55 18.91 239.74 79.12 183.08December 16.89 17.14 59.71 83.30 10.43January -35.70 19.93 43.10 25.02 2.31February 36.50 18.43 80.33 103.13 32.13

Total 6.06 844.50 726.60 1014.51 562.65

Table 2. Monthly totals for February 2002 to February 2003 for all water budget components.

265

Table 3. Sensitivity analysis for all water budget components. Percent error foreach component taken from literature.

Component Estimate % Error Error Range

mm mm mm

Precipitation 1014 5 ±50.7 963-1065

Change in Storage 6.1 5 ±0.305 5.8-6.4

Surface Outflow 727 5 ±35.4 692-762

Evapotranspiration 845 10 ±84.5 761-930

Total error 25 ±171

266

∆ Storage

Wetland

GroundwaterInflow

GroundwaterOutflow

SurfaceOutflow

PrecipET

SurfaceInflow

Figure 1. Schematic diagram representing some of the possible inputs and outputsinto a wetland. Oval represents wetland, with arrows indicating direction of watermovement.

267

Figure 2. Map of North Carolina showing location Robeson County.

268

Lumberton, NC

Juniper Bay

Figure 3. Map of Robeson County showing Juniper Bay location inrelationship to Lumberton, NC.

269

Figure 4: Map of Juniper Bay showing location of instrumentation.There is perimeter ditch and one main outflow ditch. The main outflowpoint is the location of the weirs. Circles represent location of raingauges along with the weather station near the center of the bay.

Main outflow(Location of weirs)

Perimeterditch

NW rain gauge

SE rain gauge

Weather Station

270

0

0.2

0.4

0.6

0.8

1

1.2

32 62 92 122 152 182 212 242 272 302 332 362

Day of Year

Kc

Figure 5. Crop coefficient Curve for Juniper Bay. Constructed using the ASCEHydrology Handbook procedure. Crop coefficient is used in the ETc=EToKcequation.

271

Figure 6. Map of Juniper Bay showing locations of monitoring wells.

272

0

20

40

60

80

100

120

140

160

180

200

Feb March April May June July August Sept Oct Nov Dec Jan Feb

Feb 2002-Feb 2003

mm AVG

Normal

Figure 7. Precipitation totals for each month during February 2002 to February 2003compared to normal year precipitation totals. Totals are the average of rain gauges atJuniper Bay. Bar in normal year precipitation totals is the maximum values for monthduring normal year.

273

0

20

40

60

80

100

120

140

160

Feb March April May June July August Sept Oct Nov Dec Jan Feb

Feb 2002- Feb 2003

mm

Figure 8. ET totals for each month during February 2002 to February 2003

274

Figure 9. Monthly totals for change in storage at Juniper Bay. Positive numbers indicate a rise in the water table. Negative numbers indicate a fall in the water table.

-80

-60

-40

-20

0

20

40

60

80

100

Feb

Mar

ch

Apr

il

May

June

July

Aug

ust

Sep

t

Oct

Nov

Dec Jan

Feb

Feb 2002-Feb 2003

mm

275

Figure 10. Monthly totals for surface outflow at Juniper Bay.

0

50

100

150

200

250

300

Feb

Mar

ch

Apr

il

May

June

July

Aug

ust

Sep

t

Oct

Nov

Dec Jan

Feb

Feb 2002-Feb 2003

mm

276

-50

0

50

100

150

200

Feb March April May June July August Sept Oct Nov Dec Jan Feb

Feb 2002- Feb 2003

mm

Figure 11. Monthly totals for Net Groundwater at Juniper Bay. Positive numbersindicate Juniper Bay acting as a discharge wetland. Negative numbers indicating JuniperBay acting as a recharge wetland.

277

141

563

0

50

100

150

200

250

300

350

400

450

500

550

600

50 55 60 65 70 75 80 85 90 95 100

ET(%PET)

Gi-

Go

Figure 12. Estimated Net Groundwater versus Evapotranspiration (ET) as a percent ofpotential evapotranspiration (PET). If PET was off by as much as 50 percent there isstill a positive net groundwater. Estimates of ET are needed to compute the waterbudget, but PET was estimated from meteorological data. Gi-Go is the residual of thewater budget equation and is considered the groundwater component.

278

Figure 13. USGS topographical map of Juniper Bay and surrounding area-showing elevation(in feet above sea level). Higher elevations are to east and west and lower elevations are tonorth and south.

279

Figure 14. During two-week period of no rainfall, water table indicates ET draw down during theday and replenishing of water by groundwater at night. Well 1 is located in the west, where it isthought that there is possible groundwater inflow.

Figure 15. Well 17 located in the eastern part of bay. Also has the cyclic pattern of ET draw down and groundwater inflow.

well 1

-50

-49

-48

-47

-46

-45

-44

-43

-42

-41

-40

8/1/02 8/2/02 8/3/02 8/4/02 8/5/02 8/6/02 8/7/02 8/8/02 8/9/02 8/10/02 8/11/02 8/12/02 8/13/02 8/14/02in

ches

bel

ow

su

rfac

e

well 17

-51

-50

-49

-48

-47

-46

-45

-44

-43

-42

-41

-40

-39

-38

-37

8/1/02 8/2/02 8/3/02 8/4/02 8/5/02 8/6/02 8/7/02 8/8/02 8/9/02 8/10/02 8/11/02 8/12/02 8/13/02 8/14/02

inch

es b

elo

w s

urf

ace

280

Figure 16. Well 10 located in center of bay. It does not have the cyclic pattern of ET and

groundwater inflow indicating that there is no groundwater inflow in center of bay.

well 10

-45

-40

-35

-30

-25

-20

8/1/02 8/2/02 8/3/02 8/4/02 8/5/02 8/6/02 8/7/02 8/8/02 8/9/02 8/10/02 8/11/02 8/12/02 8/13/02 8/14/02

inch

es b

elo

w s

urf

ace

281

Organic

Histic Organic

Mineral

Min

eral

Histic

Figure 17. Soil types of Juniper Bay. Areas with organic soils are possible areaswith breaks in clay layer allowing for groundwater inflow.

Area of possible Groundwater input

282

Discharge

Flowthrough

Recharge

Gi<GoGi=Go

Gi>Go

Figure 18. Topographical relationship of Juniper Bay with surrounding area. WhenGroundwater inflow is less than outflow, there is recharge of the groundwater. Whengroundwater inflow is greater than outflow, there is a discharge of surface water.Flowthrough situations occur when groundwater inflow and outflow are equal. Juniper Bayis in the intermediate elevation and should be a flowthrough wetland. However, due to theditch network Juniper Bay is acting as a discharge wetland.

283

3 mClay layer

Water Table

0 m

Main DitchPerimeter Ditch (open) Perimeter Ditch (open)

Aquifer 1

Figure 19. Schematic diagram of Juniper Bay after restoration if the perimeter ditch is leftopen. Groundwater from Aquifer 1 (Zone 1) will be intercepted by perimeter ditch andleave as surface outflow. Groundwater from Aquifer 2 (Zone 2) will seep into JuniperBay and restore wetland hydrology.

6 mBlack Creek-Regional Aquitard

Aquifer 2

284

Aquifer 2

Main DitchPerimeter Ditch Perimeter Ditch

Rim

Aquifer 1

Rim

Clay Layer

0 m

3 m

6 mBlack Creek-Regional Aquitard

Figure 20. Schematic diagram of Juniper Bay after restoration if perimeter ditch is closed.The water table might rise above the surface of the soil. Since there is no outlet for thegroundwater inflow, there could be groundwater outflow that might have offsite impacts.

285

Chapter 12

HYDROLOGIC EVALUATION OF RESTORATION OPTIONS FOR JUNIPER BAY

G.P. Fernandez, R.W. Skaggs, and M.J. Vepraskas

INTRODUCTION

Carolina Bays are found on the Atlantic Coastal Plain between northern Floridaand New Jersey (Lide et al., 1995). They are mostly heavily concentrated in easternNorth and South Carolina (Sharitz and Gibbons 1982). Most Carolina Bays are shallow,closed depressions of uncertain origin and usually contain significant areas of wetlands.Normally they function similar to an intermittent pond with water table that fluctuates inresponse to seasonal or long-term climatic conditions. Hydroperiods can range frompermanently flooded to seasonally saturated condition. The topographic gradient in thebays may induce a drainage gradient where higher elevations are excessively drained andlower elevations are poorly or very poorly drained.

The bays are generally elliptical in shape or ovate and may range from less than50 meters in length to over eight kilometers (Richardson and Gibbons 1993). Despitetheir abundance (Richardson and Gibbons (1993) estimates their numbers to vary from10,000 to 20,000), few studies have described their hydrologic regime. Recently, detailedhydrology studies have been conducted to characterize the complex hydrology ofCarolina Bays (e.g. Lide et al., 1995; O’ney et al., 1999).

This study examines the hydrology of Juniper Bay, a 705 ac Carolina Bay locatedin Robeson County in southeastern North Carolina. The main objective of the study is todescribe the hydrology of the bay. As part of an ongoing wetland mitigation project ofthe North Carolina Department of Transportation (NCDOT), this study will document thewater table fluctuations in the bay, assess the current hydrologic regime to determine ifthere are areas within the bay that are jurisdictional wetlands (satisfies the wetlandhydrologic criterion under its current condition of land use) and evaluate the variousmethodologies proposed by NCDOT to restore wetland hydrologic functions in the bay.

BACKGROUND

The North Carolina Department of Transportation (NCDOT) in compliance withSection 404 of the Clean Water Act is mandated to restore wetlands to replace thoselands that were altered or destroyed in the course of road construction and maintenance.This study is part of an ongoing wetland restoration project of NCDOT on Juniper Bay, aprior converted Carolina Bay depressional wetland in Robeson County, North Carolina.In cooperation with North Carolina State University, a research study is being conductedto evaluate the various strategies and performance of the restoration of wetland functionswithin the bay.

Wetland Determination. Land-use regulations require that frequency and duration ofsaturated conditions be determined to assess if the site satisfies jurisdictional wetland

286

criterion. Several criteria have been proposed for defining the hydrology of wetlands asregulated by Section 204 of the Clean Water Act. Among the criteria given is the 1987US Army Corp of Engineer Wetlands Delineation Manual (Environmental Laboratory,1987) which states that wetland hydrology requires an area to be saturated to the surfaceduring the growing season, for a duration between 5% and 12.5% of the growing seasonwith a frequency of at least 5 in 10 years. Saturation to the surface is assumed if the watertable is within 30 cm of the surface. Since 1992, the COE wetlands delineation manualhas been applied in making Section 404 jurisdictional wetland determinations.

Water table monitoring is commonly used to determine whether or not a sitesatisfies the wetland hydrologic criterion (Hunt et al., 2001). However, long-termmonitoring is usually needed in order make reliable assessment of the hydrology of thesite. The variability in climate requires that a longer monitoring period be used in orderfor results to be representative of normal conditions. Short-duration monitoring isinexpensive and easy to do but it is not guaranteed that the monitoring period representsthe normal climatic condition of the site. Hunt et al., (2001) have shown that decisionsbased on short-term monitoring data may be significantly in error even when weatherdata appears to be normal for the site.

An alternative approach is to use simulation models to predict water tablefluctuations over a long-term period. DRAINMOD is an example of a simulation modelthat have been applied in wetland hydrologic determinations (Skaggs and Evans, 1990;Hunt et al, 2001). When properly calibrated with site-specific data, the model accuratelypredicts the water table fluctuations in a given site as demonstrated in various fieldtesting and application (e.g. Skaggs et al., 1981; Skaggs, 1982; Chang et al., 1983; Gayleet al., 1985; Rogers, 1985; Fouss et al., 1987; Susanto et al.,1987; McMahon et al., 1988;Broadhead and Skaggs, 1989; Wright et al., 1992; Cox et al., 1994).

DRAINMOD. The field scale hydrology model, DRAINMOD, simulates water tablelevels and drainage outflow on a day-to day and hour-by-hour basis (Skaggs, 1978). Themodel requires precipitation, ET, drainage design parameters, soils and crop data asinputs. The model is generally used to simulate the performance of drainage and relatedwater table management systems over a long period of climatological data. DRAINMODhas been well tested in numerous field experiments on a wide range of soils, crops andclimatological conditions (e.g. Skaggs et al., 1981; Skaggs 1982; Fouss et al., 1987). Ithas been accepted by the USDA Natural Resource Conservation Service (USDA NRCS)as a model for design and evaluation of drainage and sub-irrigation systems in humidregions.

MATERIALS AND METHODS

Site Description

Juniper Bay is located in Robeson County, North Carolina. The bay isapproximately elliptical in shape with length of 2.4 km, width of 1.6 and an area ofapproximately 300 ha (major axis oriented NW to SE). Maximum relief is about 2 m,

287

with the perimeter of the bay at 37 m and the lowest point within the bay at 35 m abovemean sea level. The bay was developed for agriculture several decades ago. It was anintensively managed and drained agricultural land up to year 2000 when the restorationproject of NCDOT began. It is not currently a jurisdictional wetland due to its status as aprior converted agricultural land.

The organic soil in the bay is predominantly Ponzer Muck. The mineral soilsfound are Pantego fine sandy loam and Leon sand. NCDOT classified the soils in the bayinto three soil mapping units which generally corresponds to the major soil series.According to the NCDOT classification, the mapping units are a) soil unit 1 (SU-1,mostly Ponzer Muck which comprises 36% of the site, b) soil unit 2 (SU-2, mostly Leonsand which is about 50% of the site) and c) soil unit 3 (SU-3, mostly Pantego loam about13% of the site). About 1% of the area are upland soils. The bay is drained by anextensive network of drainage ditches. These ditches are from 0.6 to 1.8 m deep andspaced from about 60 m to 90 m apart. A perimeter ditch, about 4 m deep and 5 m wide,encircles the bay.

Monitoring of the existing hydrology consists of a network of water table wellsinstalled at regular intervals within and outside the bay (Fig. 1). Two well networks wereinstalled, 18 wells monitored by NCDOT (data collection starting January 2000) andanother network installed by NCSU (17 wells inside and 15 wells outside the bay). Datacollection for the wells monitored by NCSU started in March 2001. Water tables aremonitored hourly to depths of 2 m using RDS automatic monitoring wells (Remote DataSystems, Inc., Wilmington, NC). Rainfall is also measured at the bay using automaticrecording gauges. A weather station is located near the center of the bay where hourlyclimatic data (rainfall, air temperature, humidity, solar radiation and other weatherparameters) are measured and recorded.

DRAINMOD Modeling

DRAINMOD was calibrated using three years of observed data and water tablemeasurements collected at six well locations within the bay. Six sites (well locations)were selected to be representative of the major soils within the bay. The model wascalibrated separately for each site using site-specific drainage parameters (e.g. drainspacing and depth) and soil parameters. Drain spacing and drain depths for each site wereobtained from air photos and NCDOT survey. Soil water characteristics, saturatedhydraulic conductivity and other soil properties were also determined from soil coresobtained near the location of the wells. A composite rainfall data file was generated andused in the model. The composite file consisted of observed data from two recordinggauges in the bay and from the weather station at Lumberton, NC. There were gaps in therainfall data collected on the site. Missing data from the recording gauges in the bay werefilled in with the data from the Lumberton weather station. The calibration procedureinvolves the minimization of the mean absolute water table difference between theobserved and model predictions.

288

Long-term simulations (1950-2001) were conducted using the calibrated modelparameters and weather data from Lumberton, NC to characterize the hydroperiods of thesites. The hydroperiod, as defined here, is the saturation period when water table waswithin 30 cm of the surface continuously for a number of days, either at 5% or 12.5% ofthe growing season. The growing season in Robeson County is from March 14 toNovember 14 (246 days). The number of wet years (hydroperiods equal to or greater thanone) was of interest in characterizing the hydrologic status of the bay based on historicweather data. Following the hydrologic wetland criterion, a site would be considered tohave ‘wetland hydrology’ if more than half of the years qualify as ‘wet years’. That is, asite is considered to have wetland hydrology if the hydrologic criterion is satisfied inmore than 50% of the years. Evaluation of the hydrologic status prior to implementingwetland restoration methodologies is important as baseline conditions to evaluate theeffectiveness of wetland restoration.

Hydrologic restoration methods proposed for Juniper Bay includes filling thelateral ditches or placing plugs at specific intervals in the drainage ditches. Severalsurface management treatments are proposed in the NCDOT project. They includeretaining the crowned surfaces and roughing the surface to form natural depressions. Thekey question that needs to be addressed is how will the proposed treatments affect thewater table hydroperiods. In addition, effects of the treatments on soil properties, theestablishment of wetland vegetation and the desired community structure need to beaddressed.

Long-term DRAINMOD simulations (52 years) were conducted to evaluate theeffectiveness of the proposed wetland restoration methodologies. Results of thesimulations were summarized in terms of the number of wet years. The methodsevaluated are described below along with a brief description of how the treatment issimulated with DRAINMOD:

a) Ditch filling, no crown – simulated with a shallow drain depth (e.g. 5 cm)b) Ditch filling, with crown – simulated with a 30 cm drain depthc) Ditch filling, no crown, increased surface storage – simulated similar to method a but

with increase surface storage (e.g. 2.5 cm).d) Ditch filling, with crown, increased surface storage – simulated similar to method b

but with increase surface storage (e.g. 2.5 cm).

Filling the ditches without removing the field crown effectively decreases thesubsurface drainage from the site resulting in frequent shallow water tables in response torainfall events. Increasing surface storage generally decreases surface runoff and maycause water to pond on the surface. This then increases the available water for infiltrationthat will result in a consequent rise in water table by an amount dependent on thedrainable porosity of the soil. The effect of ditch filling or plugging ditches and surfacetreatments is not straightforward. The combined effect depends on soil properties anddrainage system design among others.

289

RESULTS AND DISCUSSION

Water Table Fluctuations

Figure 2 shows the water table fluctuations of the six wells used in this study forthe period from January 2000 to November 2002. The water table fluctuations aregenerally dependent on the location of the wells, the drainage system in place and thesoils of the site. Wells located on higher elevations have water table drawdowns (e.g.wells 2 and 3) as deep as 80 to 95 cm. Even during wet periods, the water table does notstay on the surface for a long time. In contrast, wells in lower elevations have longerperiods of saturation, with water tables near or at the surface for long periods, andmaximum water table drawdown 70 to 80 cm (e.g. wells 10 and 16). Water tablefluctuations of the well in soil unit 2 are similar to the responses of the wells in soil unit1. However, water tables in soil unit 1 frequently rose to the surface and stayed muchlonger. In contrast, the frequency of water tables rising to the surface in wells 2 and 3(soil unit 3) was much less than for the wells in soil units 1 and 2. Well 16 was anexception because it is located in a relatively low elevation compared to wells 2 and 3.The fluctuation in water tables is of course dependent not only on the soil properties ofthe site but also on the drainage system, as well as, surface cover and rooting depths. Asshown by the three years of data, the temporal fluctuations of the water tables within thebay are also spatially variable. The spatial distribution has to be accounted for whenimplementing wetland restoration methodologies.

DRAINMOD Simulations

DRAINMOD was calibrated with water table data for 2000 and validated with the2001-2002 data. Figures 3 to 8 show the predicted and observed water tables for six wells(wells 10 & 11 in soil unit 1, Figs. 3 & 4; well 6 in soil unit 2, Fig. 5; and wells 2, 3 & 16in soil unit 3, Figs. 6 to 8). The plots show that the model predictions are comparable toobserved water tables for all the wells. The water tables in soil unit 1 (wells 10 & 11,Figures 3 and 4) were adequately predicted. However, there are errors in predictionsduring the later part of 2001 and early 2002 where the model under predicted theobserved water tables. The rainfall events during the early part of 2002 brought theobserved water tables to the surface. During this period, the water table was within 30 cmof the surface. Although the predicted response of water table to the rainfall events wassimilar to the observed, predicted drawdown during this period was greater. This can beexplained by the installation of a control structure in late 2001 in the lateral ditch locatednear the ditches where the wells are located. The control structure resulted in higherwater levels in the ditches that reduced the subsurface drainage outflow from the fieldsnear the structure. More work on the model inputs is needed to consider this effect.

The predictions of the water tables in the wells of soil unit 2 (well 6, Fig. 5) andsoil unit 3 (wells 2, 3 and 16, Figs. 6 to 8) were much better than the predictions in soilunit 1. Except for well 16 where the control structure has a pronounced effect, predictions

290

for the other wells especially during the later part of 2001 and early 2002 were muchbetter. Wells 2 and 3 are located on a higher elevation than the other wells. Water tablesfor these wells, especially during 2001 and 2002, were much deeper than for the otherwells. The water table rarely rose to the surface in response to rainfall events.

Table 1 shows the statistics for goodness-of-fit of the predictions. In all cases,predictions of water table for all wells and for all years are within 9 cm of the measuredwater tables. Absolute error of prediction ranges from 8 cm to 13 cm. There is variabilityin prediction errors from year to year. The variability may be due to errors in rainfallmeasurements, estimation of potential ET, and spatial variability of rainfall.

Wetland Determination

Using the calibrated DRAINMOD parameters, long-term simulations (1950-2001)were conducted for each site using the weather data from Lumberton weather station.DRAINMOD simulations were conducted for each well location to characterize thehydroperiods of each site. The number of ‘wet years’ was determined. The analysis wasconducted to assess whether the sites meet the wetland hydrologic criterion under currentcondition of land use. Simulations were conducted for both the 5% and 12.5% criterion.For a growing season of 246 days (day 73 to day 318), the 5% criterion requires that thewater table be within 30 cm of the surface continuously for 13 days in 26 or more yearsof the 52 years simulation period. Similarly, the 12.5% criterion requires 31 days ofcontinuous saturation.

Table 2 summarizes results of the long-term simulations. The table shows thenumber of years that the well location satisfies the 5% or 12.5% criterion. For example, atthe 12.5% criterion, the water table at well 6 is within 30 cm of the surface continuouslyfor 31 or more days in 6 out of 52 years. Based on the 5% criterion, only well 10 and 11satisfies the hydrologic wetland criterion. Wells 10 and 11 are located in a relatively lowelevation compared to the other wells. However, if the 12.5% criterion is used, simulationresults indicate that all the well locations do not satisfy the hydrologic wetland criterionunder current condition. It is important to determine the hydrologic status of the bay priorto wetland restoration. If analysis indicates that majority of the areas within the baysatisfy the wetland hydrologic criteria, NCDOT will not get full credit for their wetlandrestoration efforts.

Evaluation of Wetland Restoration Methods

DRAINMOD simulations were conducted to evaluate various methods proposedfor restoring the hydrology of Juniper Bay site to a condition that would satisfy wetlandhydrology. Results of the simulations are summarized in Table 3. All the sites within thebay that were evaluated will satisfy the hydrologic wetland criterion (at 12.5%) if fieldditches are plugged or filled, field crowns are removed and surface storage is increased(Table 3). Filling or plugging the ditches without removing the crown will not besufficient to restore wetland hydrology for any of the sites studied. On average, thefrequency of saturation would only be around 35% for sites in the lower elevation (wells

291

10 & 11). For sites in the higher elevation such as wells 2 and 3, the frequency ofsaturation would be much lower at 4%. A dramatic increase in frequency of saturation isobtained if surface storage in the sites is increased and field crowns are not removed.Although sites in the higher elevation (wells 2 and 3) would still not satisfy thehydrologic wetland criterion at 12.5%, the frequency of saturation of the sites in thelower elevation (wells 10 and 11) increases to 78%.

If the criterion was based on water table within 30 cm of the surface for 5% of thegrowing season, ditch filling with field crowns would be sufficient to restore thehydrology of the bay to wetland conditions (Table 3, last line). The predicted frequencyof saturation ranged from 73% to 96%. These results demonstrate the importance of thecriterion on the methods that must be used to implement wetland restoration.

SUMMARY AND CONCLUSIONS

This study documents the water table fluctuations in several sites within JuniperBay, a prior converted agricultural land. Monitoring data for three years (2000-2002)indicated considerable spatial variation in the temporal patterns of the water tablefluctuations. Sites at lower elevations have water tables frequently rising to the surface orwithin the top 30 cm of the surface in response to rainfall events. Moreover, the durationof saturation is much longer. This is in contrast to sites at higher elevations. In addition totopographic differences, water tables were affected by soil properties and drainagesystem characteristics.

DRAINMOD was calibrated with observed water table data for three years andsite-specific parameters. Using the calibrated parameters, long-term simulations wereconducted to characterize the long-term variations in water table levels under currentcondition of land use and drainage characteristics. The long-term simulations indicatedthat, under current conditions, the six sites considered do not satisfy the hydrologiccriterion for wetlands if the duration is based on 12.5% of the growing season. On theother hand, if the criterion were based on 5% of the growing season (13 days), sites at thelower elevations (represented by wells 10 and 11 in the organic soils) would satisfyjurisdictional wetland status under current conditions.

Evaluation of the various restoration methodologies showed that, in order tosatisfy the 12.5% criterion for all sites, the drainage ditches would have to be filled orplugged, field crowns would have to be removed and the surface roughness increasedsuch that surface storage exceeds 2.5 cm. If the hydrologic criterion was based on 5% ofthe growing season, filling or plugging the ditches even with field crown, would besufficient to restore all the sites to wetland hydrology. In all these analyses, it wasassumed that deep or lateral seepage has insignificant impact on the water tablefluctuations.

Continued research is needed in the treatment phase of the project to determine ifthe model accurately predicted the effects of filling the ditches on water table depths and

292

wetland hydrology. Future work will consider all of sites in the bay for both pre-treatment and post-treatment conditions.

293

REFERENCES

Broadhead, R.G. and R.W. Skaggs. 1989. A hydrologic model for artificially drainedNorth Carolina peatlands, p61-70. In:V.A. Dodd and P.M. Grace (eds.)Agricultural Engineering, Proc of the 11th International Congress, Dublin, Sept 4-8.

Chang, A.C., R.W. Skaggs, L.F. Hermsmeier, and W.R. Johnson. 1983. Evaluation of awater management model for irrigated agriculture. Trans of ASAE 26:412-418.

Cox, J.W., D.J. McFarlane and R.W. Skaggs. 1994. Field evaluation of DRAINMOD forpredicting waterlogging intensity and drain performance in South-WesternAustralia. Austral J. Soil Res. 32:653-671.

Environmental Laboratory. 1987. Corps of Engineers Wetland Delineation Manual.Technical Report Y-81-1, US Army Corp of Engineer Waterways ExperimentStation, Vicksburg, MS 100p.

Fouss, J.L. R.L. Bengston and C.E. Carter. 1987. Simulating subsurface drainage in theLower Mississippi Valley with DRAINMOD. Trans of ASAE 30(6):1679-1688.

Gayle, G.A., R.W. Skaggs and C.E. Carter. 1985. Evaluation of a water managementmodel for a Louisiana sugar cane field. J.Am.Soc. Sugar-Cane Technol. 4:18-28.

Hunt, W.F., R.W. Skaggs, G.M. Chescheir and D.M. Amatya. 2001. Examination of thewetland hydrologic criterion and its application in the determination of wetlandhydrologic status. Technical report No 333. Water Resources Research Institute,Raleigh, NC.

Lide, R.F., V.G. Meentemeyer, J.E. Pinder III and L.M. Beatty. 1995. Hydrology of aCarolina Bay located on the upper Coastal Plain of western South Carolina.Wetlands 15(1):47-57.

McMahon, P.C., S. Mostaghimi, and F.S. Wright. 1988. Simulation of corn yield by awater management model for a coastal plain soil. Trans ASAE 31:734-742.

O’ney, S.E., M.H. Eisenbies and M.Miwa. 1999. Hydrologic processes in the vicinity of aCarolina bay affecting water quality: an assessment in association with ahardwood fiber farm. Unpublished progress report. USDA Forest Service, Centerfor Forested Wetlands Research, Charleston, SC.

Richardson, C.J. and J.W. Gibbons. 1993. Pocosins, Carolina Bays, and Mountain Bogs.P 257-309. In W.H. Martin, S.G. Boyce and A.C. Echternacht (eds). Biodiversityof the Southeastern United States/Lowland Terrestial Communities. John Wileyand Sons, Inc. New York, NY, USA.

294

Rogers, J.S. 1985. Water management model evaluation for shallow sandy soils. TransASAE 28:785-790.

Sharitz, R.R. and J.W. Gibbons. 1982. The ecology of southeastern shrub bogs (pocosins)and Carolina Bays: a community profile. FWS/OBS-82/04. Division of BiologicalSciences, US Fish and Wildlife Service, Washington DC.

Skaggs, R.W. 1978. A water management model for shallow water tables soils. TechnicalReport 134. Water Resources Research Institute, Raleigh, N.C.

Skaggs, R.W. 1982. Field evaluation of a water management model. Trans of ASAE25:666-674.

Skaggs, R.W., W.R. Fausey and H.B. Nolte. 1981. Water management model evaluationfor North Central Ohio. Trans of ASAE. Vol 24(4):922-928.

Skaggs, R.W., and R.O. Evans. 1990. Methods to evaluate effects of drainage on wetlandhydrology. In: J.A. Kusler and S. Daly (eds.) Wetlands and River CorridorManagement, Association of Wetland Managers, Berne, NY:291-299.

Susanto, R.J., J.Feye, W.Dierickx and G.Wyseure. 1987. The use of simulation models toevaluate the performance of subsurface drainage systems. p.A67-A76. In Proc 3rd

Int. Drain. Workshop on Land Drain. Columbus, OH. Ohio Dept of Agric Eng.,Ohio State Univ., Columbus.

Wright, J.A., A. Shirmohammadi, W.L. Magete, J.L. Fouss, R.L. Broughton and J.E.Parsons. 1992. Water table management practice effects on water quality. TransASAE 35:823-831.

295

Table 1. Statistics for goodness-of-fit of the predicted water tables for 2000-2002.

Well 2 Well 3 Well 6 Well 10 Well 11 Well 16Error, cm

2000 1.8 5.0 -1.4 -9.4 -1.6 -4.22001 -11.6 -2.4 1.2 -7.7 -7.7 -11.92002 2.1 1.2 -1.7 -5.2 -7.6 -8.22000-02 -3.0 1.1 -0.6 -7.4 -5.6 -8.3

Absolute Error, cm2000 6.6 8.6 6.6 12.2 7.2 8.52001 15.8 9.2 8.5 10.1 10.1 17.02002 7.3 6.6 9.5 12.2 14.0 15.22000-02 10.2 8.2 8.2 11.5 10.5 13.8

Table 2. Number and percentage of wet years in 52 years of simulation.

Soil Unit 3 Soil 2 Soil Unit 1

Well 2 Well 3 Well 16 Well 6 Well 10 Well 11

5 % (13 days) 5(10%)

6(12%)

20(38%)

21(40%)

37(71%)

37(71%)

12.5% (31 days) 0 0 0 0 6(12%)

4(8%)

Table 3. Number and percentage of wet years in 52 years of simulation (12.5% or 31 dayssaturation).

Soil Unit3

Soil 2 Soil Unit1

Treatment Well2

Well3

Well16

Well6

Well10

Well11

Ditch Fill, with crown 2(4%)

2(4%)

2(4%)

6(12%)

19(33%)

17(37%)

Ditch Fill, no crown 10(19%)

10(19%)

3(6%)

10(19%)

22(42%)

23(44%)

Ditch Fill, with crown, surfacestorage = 2.5 cm

17(33%)

17(33%)

23(42%)

20(38%)

40(77%)

41(78%)

Ditch Fill, no crown, surfacestorage = 2.5 cm

31(60%)

31(60%)

32(62%)

33(63%)

43(83%)

45(87%)

Ditch Fill, with crown, at 5% (13days)

42(81%)

38(73%)

38(73%)

43(83%)

50(96%)

50(96%)

296

Figure 1. Juniper Bay with the network of monitoring wells and soil core locations. TheD labeled wells were used in the study.

#Y

#Y

#Y

#Y

#Y#Y

#Y

#Y

#Y

#Y#Y

#Y

#Y#Y

#Y

#Y

#Y

#Y

#Y

#Y

#Y#Y

$

$

$

$$

$$$

$

$$

$$ $$

$

$

$

$$$

$$

$

$$$

$

$$

$

$

$

$$$$$$

$

$

$

D2

D16

D6

D11

D3

D10

RainGage

Project WellsSoil CoresNCDOT Wells

297

Figure 2. Water table fluctuations from January 2000 to November 2002 at six wells inJuniper Bay.

- 9 0

- 8 0

- 7 0

- 6 0

- 5 0

- 4 0

- 3 0

- 2 0

- 1 0

0

1 0

2 0

2 0 0 0 2 0 0 0 . 5 2 0 0 1 2 0 0 1 . 5 2 0 0 2 2 0 0 2 . 5 2 0 0 3

Wat

er T

able

Dep

th,

cm

D O T W e l l 1 1 D O T W e l l 1 0S o i l U n i t 1

- 9 0

- 8 0

- 7 0

- 6 0

- 5 0

- 4 0

- 3 0

- 2 0

- 1 0

0

1 0

2 0 0 0 . 0 2 0 0 0 . 5 2 0 0 1 . 0 2 0 0 1 . 5 2 0 0 2 . 0 2 0 0 2 . 5 2 0 0 3 . 0

Wat

er T

able

Dep

th,

cm

D O T W e l l 6S o i l U n i t 2

- 1 0 0

- 9 0

- 8 0

- 7 0

- 6 0

- 5 0

- 4 0

- 3 0

- 2 0

- 1 0

0

1 0

2 0

2000.0 2 0 0 0 . 5 2 0 0 1 . 0 2 0 0 1 . 5 2002.0 2 0 0 2 . 5 2 0 0 3 . 0

Wat

er T

able

Dep

th, c

m

D O T W e l l 3 D O T W e l l 2 D O T W e l l 1 6S o i l U n i t 3

298

Figure 3. Observed and predicted water table depth at well 10 in Soil Unit 1.

- 9 0

- 8 0

- 7 0

- 6 0

- 5 0

- 4 0

- 3 0

- 2 0

- 1 0

0

1 0

2 0 0 2 . 0 2 0 0 2 . 1 2 0 0 2 . 2 2 0 0 2 . 3 2 0 0 2 . 4 2 0 0 2 . 5 2 0 0 2 . 6 2 0 0 2 . 7 2 0 0 2 . 8 2002.9 2003.0

Wat

er T

able

Dep

th, c

m

D O T W e l l 1 0 Predicted

-90

-80

-70

-60

-50

-40

-30

-20

-10

0

1 0

2000 .2 2000 .3 2000.4 2000.5 2000.6 2000.7 2000 .8 2000 .9 2001 .0W

ater

Tab

le D

epth

, cm

D O T W e l l 1 0 Pred ic ted

-90

-80

-70

-60

-50

-40

-30

-20

-10

0

1 0

2001 .0 2001 .1 2001 .2 2001 .3 2001 .4 2001 .5 2001 .6 2001 .7 2001 .8 2001.9 2002.0

Wat

er T

able

Dep

th, c

m

D O T W e l l 1 0 Predicted

299

Figure 4. Observed and predicted water table depth at well 11 in Soil Unit 1.

-90

-80

-70

-60

-50

-40

-30

-20

-10

0

1 0

2000 .2 2000.3 2000.4 2000 .5 2000 .6 2000.7 2000.8 2000 .9 2001 .0W

ater

Tab

le D

epth

, cm

DOT Wel l 11 Predicted

- 9 0

- 8 0

- 7 0

- 6 0

- 5 0

- 4 0

- 3 0

- 2 0

- 1 0

0

1 0

2 0 0 1 . 0 2 0 0 1 . 1 2 0 0 1 . 2 2001 .3 2001 .4 2001 .5 2001 .6 2001 .7 2001 .8 2001.9 2002.0

Wat

er T

able

Dep

th, c

m

D O T W e l l 1 1 Predic ted

- 9 0

- 8 0

- 7 0

- 6 0

- 5 0

- 4 0

- 3 0

- 2 0

- 1 0

0

1 0

2 0 0 2 . 0 2 0 0 2 . 1 2 0 0 2 . 2 2002 .3 2002 .4 2002 .5 2002 .6 2002 .7 2002 .8 2002.9 2003.0

Wat

er T

able

Dep

th, c

m

D O T W e l l 1 1 Predic ted

300

Figure 5. Observed and predicted water table depth at well 6 in Soil Unit 2.

-90

-80

-70

-60

-50

-40

-30

-20

-10

0

10

2 0 0 0 . 2 2 0 0 0 . 3 2 0 0 0 . 4 2 0 0 0 . 5 2000.6 2000.7 2000.8 2000.9 2001.0

Wat

er T

able

Dep

th, c

m

D O T W e l l 6 Pred ic ted

- 1 1 0

- 9 0

- 7 0

- 5 0

- 3 0

- 1 0

1 0

2001.0 2 0 0 1 . 1 2 0 0 1 . 2 2 0 0 1 . 3 2 0 0 1 . 4 2001.5 2 0 0 1 . 6 2 0 0 1 . 7 2 0 0 1 . 8 2 0 0 1 . 9 2002.0

Wat

er T

able

Dep

th, c

m

D O T W e l l 6 P r e d i c t e d

- 1 1 0

- 9 0

- 7 0

- 5 0

- 3 0

- 1 0

1 0

2002.0 2 0 0 2 . 1 2 0 0 2 . 2 2 0 0 2 . 3 2 0 0 2 . 4 2002.5 2 0 0 2 . 6 2 0 0 2 . 7 2 0 0 2 . 8 2 0 0 2 . 9 2003.0

Wat

er T

able

Dep

th, c

m

D O T W e l l 6 P r e d i c t e d

301

-100

-90

-80

-70

-60

-50

-40

-30

-20

-10

0

1 0

2000.2 2000 .3 2000.4 2000.5 2000 .6 2000.7 2000 .8 2000 .9 2001.0

Wat

er T

able

Dep

th, c

m

DOT Wel l 2 Predicted

-100

-90

-80

-70

-60

-50

-40

-30

-20

-10

0

1 0

2001.0 2001 .1 2001.2 2001 .3 2001.4 2001 .5 2001.6 2001 .7 2001.8 2001 .9 2002.0

Wat

er T

able

Dep

th, c

m

DOT Wel l 2 Predicted

-100

-90

-80

-70

-60

-50

-40

-30

-20

-10

0

1 0

2002.0 2002 .1 2002.2 2002 .3 2002.4 2002 .5 2002.6 2002 .7 2002.8 2002 .9 2003.0

Wat

er T

able

Dep

th, c

m DOT Wel l 2 Predicted

Figure 6. Observed and predicted water table depth at well 2 in Soil Unit 3.

302

-100

- 9 0

- 8 0

- 7 0

- 6 0

- 5 0

- 4 0

- 3 0

- 2 0

- 1 0

0

1 0

2000.2 2000 .3 2 0 0 0 . 4 2 0 0 0 . 5 2000 .6 2 0 0 0 . 7 2 0 0 0 . 8 2000.9 2001 .0W

ater

Tab

le D

epth

, cm

D O T W e l l 3 Predic ted

-100

- 9 0

- 8 0

- 7 0

- 6 0

- 5 0

- 4 0

- 3 0

- 2 0

- 1 0

0

1 0

2001.0 2001 .1 2 0 0 1 . 2 2001 .3 2 0 0 1 . 4 2001.5 2001 .6 2 0 0 1 . 7 2001 .8 2 0 0 1 . 9 2002.0

Wat

er T

able

Dep

th, c

m

DOT We l l 3 Predic ted

-100

-90

-80

-70

-60

-50

-40

-30

-20

-10

0

1 0

2002.0 2002.1 2002 .2 2002.3 2002 .4 2002.5 2002.6 2002 .7 2002.8 2002 .9 2003.0

Wat

er T

able

Dep

th, c

m DOT Wel l 3 Predicted

Figure 7. Observed and predicted water table depth at well 3 in Soil Unit 3.

303

Figure 8. Observed and predicted water table depth at well 16 in Soil Unit 3.

-90

-80

-70

-60

-50

-40

-30

-20

-10

0

10

2 0 0 0 . 2 2 0 0 0 . 3 2 0 0 0 . 4 2 0 0 0 . 5 2 0 0 0 . 6 2 0 0 0 . 7 2 0 0 0 . 8 2 0 0 0 . 9 2 0 0 1 . 0

Wat

er T

able

Dep

th, c

m

D O T W e l l 1 6 P r e d i c t e d

- 9 0

- 8 0

- 7 0

- 6 0

- 5 0

- 4 0

- 3 0

- 2 0

- 1 0

0

1 0

2 0 0 1 . 0 2 0 0 1 . 1 2 0 0 1 . 2 2 0 0 1 . 3 2 0 0 1 . 4 2 0 0 1 . 5 2 0 0 1 . 6 2 0 0 1 . 7 2 0 0 1 . 8 2 0 0 1 . 9 2 0 0 2 . 0

Wat

er T

able

Dep

th, c

m

D O T W e l l 1 6 P r e d i c t e d

- 9 0

- 8 0

- 7 0

- 6 0

- 5 0

- 4 0

- 3 0

- 2 0

- 1 0

0

1 0

2 0 0 2 . 0 2 0 0 2 . 1 2 0 0 2 . 2 2 0 0 2 . 3 2 0 0 2 . 4 2 0 0 2 . 5 2 0 0 2 . 6 2 0 0 2 . 7 2 0 0 2 . 8 2 0 0 2 . 9 2 0 0 3 . 0

Wat

er T

able

Dep

th, c

m

D O T W e l l 1 6 P r e d i c t e d

304

Chapter 13

PRINCIPAL FINDINGS AND CONCLUSIONS

M.J. Vepraskas

1. Juniper Bay was formed in sediments that comprise the Middle Atlantic Coastal Plain.Borings ranging to depths of 20 to 50 ft showed the sediments below the bay consist ofalternating layers of clay and sand of varying thicknesses. Lateral variability in theconcentrations of clay is considerable above a depth of approximately 20 ft. Below 20 ft., acontinuous clay layer may extend beneath the bay and the surrounding area. This sedimenthad a characteristic sequence of layers which were found both within and outside the bay.Clay layers within 20 ft. of the surface may not be continuous beneath the bay, in partbecause the sediments appear to have been cut by river or stream channels that have filledin.

2. A compilation of the history of land use at Juniper Bay showed that the upper 3 ft of soilhas been extensively modified by activities extending back to 1909. Major landmodifications have occurred since the land was prepared for agriculture, beginning in 1979.The northwest section of the Bay has had two different ditch networks placed within it, thefirst one having been filled in 1981 when the current ditch network was constructed. Whileclearing the land for agriculture, debris was burned on site. The organic soil did catch fireand 2 to 3 ft of the soil was burned off, possibly over a 2-year period. A small lake formedin the depression until filled in. Agricultural chemicals have been applied to the soilannually since 1979.

3. Vegetation surveys were conducted in three references bays found in Bladen County. Fourplant communities were identified: pond pine woodland, non- riverine swamp forest, bayforest, and high pocosin. Of these types, only the pond pine woodland seems appropriatefor Juniper Bay. Non- riverine swamp forest was found in only one bay and its hydrologicrequirements are unknown at this time. Bay forests and high pocosins were found on deeporganic soils, which do not occur at this time in Juniper Bay.

4. Annual chemical additions over the last 30 years have raised the levels of Ca, Mg, P, and Kin the soils of Juniper Bay to a depth of 30 in. Soils in reference bays had minisculequantities of plant nutrients, particularly below the upper 12 in., as compared to JuniperBay. The effect of the increased nutrient levels on native plant establishment is unknown atthis time.

5. Soil physical properties have changed the most in Juniper Bay’s organic soils followingbeing placed into agricultural production. The organic soils have become aggregated in theupper 12 in. as compared to similar soils in the reference bays. The effect of this on plantestablishment of native species will probably be minor.

6. A procedure was developed to estimate the amounts of subsidence the organic soils haveundergone in Juniper Bay over the last 30 years. The results showed that the organic

305

surface is approximately 24 in. lower today than before the Bay was cleared. Consideringthat the restored water table is generally at or near the top of the organic surface during thecourse of a year, this finding suggests that if the ditches are completely plugged at JuniperBay the water will rise to a level 24 in. above the present surface.

7. Surface outflow nutrient levels in Juniper Bay were generally low, and probably too low tocause eutrophication in downstream waters at this time. There was little difference betweennutrient concentrations in surface water outflow between storms, and that leaving the Bay instorms whose intensity was less than 2 in./day. However, for rain events greater than 2in./day there was a two-fold increase in nutrient concentrations in surface outflow. Thisincrease can be attributed to the heavy rain that exceeded the infiltration rate of the soil andcaused surface runoff. These nutrient levels stay elevated for about 1 week because of thetime it took for ditches in the entire Bay to drain. There could be short-term water qualitydegradation with these storm events but long-term eutrophication should not be a problem.It should be emphasized that these measurements pertain to pre-restoration conditions.

8. Ground penetrating radar (GPR) surveys showed that the shallow aquitard (clay layer)depth ranged from 20 in. to 10 ft. with an average depth of 60 in. across the site. One ormore aquitards are present below Juniper Bay, with aquitards frequently overlying eachother. However, the GPR survey detected an anomaly in the southeast corner of JuniperBay that where no aquitard could be detected. Coring in this region confirmed that noclayey soil horizons existed within 15 ft of the surface. The size of the anomaly isestimated to be 8 acres, or approximately 1% the area of Juniper Bay. In addition, it islikely that at least the main N-S ditch has pierced underlying aquitards. Any water thatmight have been retained near the surface by these aquitards could be drained off site.NCDOT may want to consider a clay or synthetic lining in the base of the main ditch orother ditches to prevent such water loss.

9. Measurements of hydraulic gradients indicated that groundwater is moving below JuniperBay in three zones or aquifers, as shown below, which are separated by one or more claylayers.

Bottom Clay Layer

Perimeter ditch Main outflow ditch Lateral ditch

“Middle” Clay Layer

2.44 - 3.05 m (8 - 10 ft)

3.05 - 3.66 m (10 - 12 ft) Zone 1

Zone 2

Zone 3

306

Zone 2 may consist of alternating clay and sand layers. Hydraulic gradient measurementsshowed that groundwater was entering Juniper Bay through zone 2 in the western andeastern sides of the bay. Groundwater was discharging to the north and south.Groundwater moving through zone 2 appeared to be moving into zone 1 in the northwesternportion of the bay. Groundwater flowing through zone 3 moved to the southeast and didnot appear to move into the overlying zones. Groundwater moving in zone 1 flowed to theditches and was removed as surface water.

10. The water budget showed that net groundwater input into Juniper Bay may comprisebetween 10 and 35% of the total water input into the Bay. The wide variation in theestimate is due to uncertainty in the actual value of evapotranspiration. Fluctuation in welllevels and hydraulic gradients support a conclusion that some groundwater is entering theBay from the east and west. A portion of the groundwater is probably leaving the site assurface water.

11. Modeling of the current hydrology using DRAINMOD showed that under currentconditions, the six sites considered do not satisfy the hydrologic criterion for wetlands if theduration is based on 12.5% of the growing season. On the other hand, if the criterion werebased on 5% of the growing season (13 days), sites at the lower elevations (represented bywells 10 and 11 in the organic soils) would satisfy jurisdictional wetland status undercurrent conditions.

Evaluation of the various restoration methodologies showed that, in order to satisfy the12.5% criterion for all sites, the drainage ditches would have to be filled or plugged, fieldcrowns would have to be removed and the surface roughness increased such that surfacestorage exceeds 2.5 cm. If the hydrologic criterion was based on 5% of the growing season,filling or plugging the ditches even with field crown, would be sufficient to restore all thesites to wetland hydrology. In all these analyses, it was assumed that deep or lateral seepagehas insignificant impact on the water table fluctuations.

307

Chapter 14

RECOMMENDATIONS AND TECHNOLOGY TRANSFER PLAN

RECOMMENDATIONS

1. Further research is needed to determine the amount of ground water entering Juniper Bay.Until we fully understand the impact of the groundwater on the Bay’s hydrology, theperimeter ditch should be left open and it should be maintained to keep water flowing.Beaver dams need to be removed periodically, the beaver population controlled.

2. Soils have unnaturally high fertility levels near the surface. Removal of more than 24 in. ofsoil material will expose less fertile subsoil material. This will have an as yet unknownimpact on planted vegetation. Modifications to the soil surface should be kept to theminimum necessary to reestablish the natural hydrology. All surfaces should be covered intopsoil after modification.

3. Once hydrology is restored and the soils become anaerobic, the concentration of P in theground water and surface outflow through the perimeter ditch will increase and may exceedacceptable levels. NCDOT needs to evaluate the potential regulatory problems this couldpose.

TECHNOLOGY TRANSFER

The data collected to date comprise one of the largest sets of information on Carolina Bays in thecountry. This information will be disseminated through research articles, talks at professionalmeetings, and field trips. Chapters 2 through 12 of this report each represent a scientific article thatis being prepared for publication. Information in each of the chapters has or will be presented at ascientific meeting.

Beginning in 2004, we plan to conduct field tours at Juniper Bay to illustrate the soil properties,and hydrology of Carolina Bays. Field tours will be conducted annually for the foreseeable future.